Language or Dialect

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 25/8/2020, SPi
Language or Dialect?
Language or Dialect?
The History of a Conceptual Pair
RAF VAN ROOY
1
OUP UNCORRECTED AUTOPAGE PROOFS – FIRST PROOF, 13/11/2023, SPi
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Raf Van Rooy 2020
The moral rights of the author have been asserted
First Edition published in 2020
Some rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, for commercial purposes,
without the prior permission in writing of Oxford University Press, or as expressly
permitted by law, by licence or under terms agreed with the appropriate
reprographics rights organization.
This is an open access publication, available online and distributed under the terms of a
Creative Commons Attribution - Non Commercial - No Derivatives 4.0
International licence (CC BY-NC-ND 4.0), a copy of which is available at
http://creativecommons.org/licenses/by-nc-nd/4.0/.
Enquiries concerning reproduction outside the scope of this licence
should be sent to the Rights Department, Oxford University Press, at the address above
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 0000000000
ISBN 978–0–19–884571–3
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
For Elien
[ksin ε ˈɣεrə]
Contents
Acknowledgements xiii
List of Figures xv
List of Tables xvii
Conventions xix
1. Introduction 1
I. PREHISTORY, 500 –1500

2. A dive into the prehistory of the conceptual pair 15
2.1 The Greek dialects between anecdotes and definitions 16
2.2 Diálektos, a variety of interpretations 18
2.3 Philology as stimulus 21
2.4 The Latin West 22
2.5 Conclusion 27
3. The exception to the rule: Lingua and idioma in
Roger Bacon’s thought 28
3.1 Bacon at Babel 30
3.2 Another first for Bacon? 31
3.3 An outlook on different languages and their dialects:
a central precondition 35
3.4 Thomas Aquinas: a like-minded exegete? 38
3.5 The Tower of Babel 40
3.6 Conclusion 42
II. THE ORIGIN OF THE CONCEPTUAL

PAIR, 1500–50
4. From dogs and hounds to languages and dialects: The conceptual
pair in Conrad Gessner’s work 47
4.1 Exploring the linguistic world 49
4.2 Conrad Gessner, certified cataloguer and compiler 52
4.3 Ad fontes! Gessner reads Clement of Alexandria 55
4.4 The meanings of dialectus according to Gessner 57
4.5 Conclusion 61
viii 
5. Lingua and dialectus: From synonymy to contrast 62

5.1 Dialectus as a Latin word 62
5.2 The key symptom: contrasting lingua to dialectus 64
5.3 Developing the concept of common language 67
5.4 Updating dialectus definitions 72
5.5 The coining of a new phrase: ‘to differ only in dialect’ 74
5.6 An intricate conceptual web 76
5.7 Conclusion 79
6. Hellenism, standardization, and info-lust: The genesis of
the conceptual pair in context 81
6.1 A major pivoting point: the rediscovery of the Greek dialects 81
6.2 From ancient Greece to Western Europe: standardizing
the vernaculars 84
6.3 Knowledge revolution: information explosion and info-lust 87
6.4 The countability of language 88
6.5 A product of appropriation and subconscious adaptation 90
6.6 Conclusion 91
III. CONSOLIDATION BY ELABORATION,

1550–1650
7. Space and nation: Greek definitions transformed 95
7.1 The spatial conception of dialect 95
7.2 Two topical topoi 98
7.3 Regional language variation: a universal phenomenon? 99
7.4 Nation or nations? The ethnic conception of the language/dialect
distinction 101
7.5 Towards a political interpretation of ‘nation’? 105
7.6 Space and nation: diagnostic criteria? 107
7.7 Conclusion 108
8. Aristotle’s legacy: Substance, accidents, and mutual intelligibility 109
8.1 Extending the ‘to differ only in dialect’ phrase 109
8.2 Mithridates: polyglot or not? 113
8.3 A question of gradation: devising different levels of dialects 114
8.4 Mutual intelligibility: an early modern criterion 117
8.5 Johannes Goropius Becanus and immediate mutual intelligibility 119
8.6 The communicative reach of dialects versus languages 122
8.7 Conclusion 123
9. A subjective touch: Language beats dialect 125
9.1 Analogical norm or anomalous deviation? 125
9.2 From common to standard language 127
9.3 Superior or inferior? 129
9.4 Conclusion 134
 ix
10. The conceptual pair and language history: Language

generates dialects 136
10.1 Language mothers and offshoot dialects 136
10.2 An early critic of the language-historical interpretation 141
10.3 A discursive strategy for historical classification 143
10.4 Conclusion 146
11. Consolidation by elaboration: Drawing the balance 147
11.1 The seven main interpretations: a synthesis 147
11.2 Emancipating the conceptual pair from the Greek heritage 151
11.3 Conclusion 156
12. The conceptual pair in transition: The case of Georg Stiernhielm 159
12.1 Georg Stiernhielm, a Swedish all-round scholar 159
12.2 The conceptual pair according to Stiernhielm 163
12.3 Conclusion 166
IV. SYSTEMATIZATION AND RATIONALIZATION,

1650–1800
13. Putting the conceptual pair on the scholarly agenda:
The orientalist Albert Schultens 171
13.1 Schultens’s definitions of dialect 174
13.2 Language, dialect, and degenerate offshoot 176
13.3 Classes of linguistic variation 178
13.4 Schultens’s legacy 180
13.5 Conclusion 181
14. Lexicostatistics avant la lettre: The historian Johann Christoph
Gatterer and the conceptual pair 183
14.1 Gatterer’s Vorrede 184
14.2 Characteristic words: Gatterer and basic vocabulary 186
14.3 Determining the degree of linguistic kinship: Gatterer’s
lexicostatistic framework 187
14.4 The historian Gatterer and the grammarian Adelung 190
14.5 Conclusion 191
15. Classes of variation: How do languages and dialects differ? 193
15.1 Casting off John the Grammarian’s yoke 193
15.2 From scattered comments to systematization 195
15.3 Two eighteenth-century outsiders in the quest for
linguistic criteria 200
15.4 Conclusion 202
16. Between systematization and rationalization: The conceptual
pair through the Enlightenment lens 204
16.1 Towards a dialectological tradition? 204
16.2 The first sceptical voices 211
16.3 Conclusion 216
x 
V. FROM SILENT ADOPTION TO OUTSPOKEN

ABANDONMENT, AFTER 1800
17. From Jones to Gabelentz: Silent adoption and renewed suspicion 219
17.1 The beginnings of modern linguistics 220
17.2 Two pioneering dialectologists: Johann Andreas Schmeller
and Albert Giese 221
17.3 Questioning the conceptual pair: August Schleicher
and William Dwight Whitney 225
17.4 The late nineteenth century: between continued usage
and increasing scepticism 228
17.5 Conclusion 229
18. Schuchardt the iconoclast 231
18.1 Tree or waves? 231
18.2 A pioneering outsider 233
18.3 Breaking down the walls between languages and dialects 235
18.4 Linguists, shibbolethists 238
18.5 Useless abstractions? 240
18.6 Conclusion 242
19. From Saussure to 1954: Structuralism and the language/dialect
distinction 244
19.1 In Saussure’s class 245
19.2 Towards a structural dialectology: Uriel Weinreich’s
diasystem 247
19.3 Redefining the conceptual pair: Martinet and Polák 251
19.4 Conclusion 254
20. Mutual intelligibility: The number one criterion? 256
20.1 A criterion making career 258
20.2 Measuring linguistic distance 259
20.3 Conclusion 261
21. Between two extremes: Generative and sociolinguistic
interpretations 263
21.1 Generative grammar: no country for old dialects? 263
21.2 Variables over systems 266
21.3 Destructive dismissal 268
21.4 Supplementing the conceptual pair 270
21.5 By way of conclusion: a quest for alternatives 273
22. A gentle goodbye? Dialect stripped for parts 275
22.1 Farewell to the conceptual pair? 278
22.2 The language/dialect distinction after 1900: the story
of a love-hate relationship 280
 xi
23. Language, dialect, and the general public—or how not

to popularize knowledge 284
23.1 The conceptual pair popularized 284
23.2 The Weinreich witticism in context: the nation-state,
language policy, and mass education 288
23.3 Political activations of the conceptual pair 289
23.4 Conclusion 293
24. Language and dialect between past and future: Terminological
success, conceptual failure? 296
References 303
Index 333
Acknowledgements
This book presents ideas which I first began to develop during my PhD fellowship
at KU Leuven in the years 2013–17. This research led to a dissertation which
I defended in late May 2017 under the supervision of Toon Van Hal and Pierre
Swiggers, to both of whom I owe a great deal. It was their fascinating work which
excited my interest in the history of linguistics as a field of research. Their
feedback on my dissertation has proved invaluable in preparing the present
monograph, in which I have tried to decompress my ideas and to bridge the gap
between 1800 and the present, left open in my dissertation. I have been able to do
so by carrying out extra research during the Winter semester of 2018–19 in
Athens. Both my PhD fellowship and my stay in Greece were made possible by
the generous support of the Research Foundation—Flanders (FWO).
I am likewise much indebted to numerous other scholars, many of whom have
selflessly acted as true mentors. To this category belong first and foremost John
Considine, Lambert Isebaert, and Han Lamers. They have, moreover, shared with
me countless invaluable documents, both primary sources and secondary litera-
ture. For the same reason, Josef Eskhult deserves a specific mention here, since he
was so generous as to grant me access to his work in progress and to send me scans
of source material related to Georg Stiernhielm. I also thank Alexander Maxwell
for his helpful and very detailed feedback on Chapters 7–16.
I owe a deep debt of gratitude to the staff of libraries in Amsterdam, Athens,
Berlin, Copenhagen, Edmonton, Gera, Ghent, Jena, Lille, Leuven, London,
Madrid, Munich, Paris, Rostock, and Vienna. I am particularly grateful to the
staff of the Forschungsbibliothek in Gotha and the University Library in Leiden,
where I have been able to work in perfect peace, generously supported by a
Herzog-Ernst-Stipendium of the Fritz Thyssen Stiftung and by a Scaliger
Fellowship of the Leiden Scaliger Institute, respectively.
I am also grateful to the people behind the many and often freely available
digital initiatives, not in the least Google Books and its partnerships with institu-
tions like the Bayerische Staatsbibliothek, without which my research would have
been unfeasible. I have, moreover, had privileged access to useful databases such as
Early English Books Online and Eighteenth Century Collections Online thanks to,
among other institutions, KU Leuven and the University of Alberta. It was, as a
result, often unnecessary to spend much time travelling to research libraries across
Europe and beyond, even if there are certain dangers tied to the dematerialization
of the printed book (Grafton 2009). It may, for instance, lead to a neglect of
important material aspects of the books digitized, such as their size, or to a neglect
xiv 
of individual copies and their unique characteristics, since often only one copy of a
book is available in digital format. Additionally, maps and schemes that are
printed on pages of larger size are sometimes omitted in online reproductions.
Be that as it may, the advantages outweigh the disadvantages, and I am grateful
that I have been able to make optimal use of these digital resources.
Optical character recognition (OCR) software has been an indispensable tool,
too, as this made it possible to perform goal-oriented searches in digitized text files
in Roman script, even though it must be added that, in its current state, the tool is
not infallible, certainly not for early modern books. It is, however, my hope and
expectation that, in the future, the technique will greatly improve and will be
extended to non-Roman characters.
Last but not least, my wife, Elien, and my parents have as always supported me
unconditionally in writing and finalizing this book; for this, and for everything
else, I admire and love them.
List of Figures
3.1. Roger Bacon at his Merton College observatory, oil painting by Ernest Board 29
3.2. The Tower of Babel through the eyes of Pieter Bruegel the Elder, 1563 41
4.1. Dogs in Conrad Gessner’s work 48
4.2. World map by Abraham Ortelius (1527–98), 1570 50
4.3. Conrad Gessner, portrait by Tobias Stimmer (1539–84) 53
4.4. Marble bust of the Pontic king Mithridates VI, first century  54
5.1. Conceptualization of a language (e.g. Greek) before the early Cinquecento 64
5.2. Conceptualization of a language (e.g. Greek) emerging in the early
Cinquecento 64
5.3. Desiderius Erasmus holding a Greek book, by Hans Holbein, 1523 65
5.4. Antesignanus’s perception of the relationship between Greek and
French variation 71
6.1. The first separate edition of Adrien Amerot’s On the diverse dialects, 1530 83
6.2. Woodcut showing sixteenth-century printing in action, 1568 86
6.3. Natural history cabinet, 1599 87
7.1. Claude de Saumaise (1588–1653) 104
8.1. Bust of Aristotle, after an original of about 330  110
8.2. Johannes Goropius Becanus by Philips Galle, 1572 119
10.1. Joseph Justus Scaliger’s influential Diatribe on the languages of the Europeans 137
10.2. Scaliger c. 1607, engraving by Johann Theodor de Bry (1561–1623) 138
10.3. Johann Heinrich Hottinger etched by Georg Meyer, 1664 142
11.1. The dimensional triangle shaping the early modern conceptual pair 149
12.1. Georg Stiernhielm 160
12.2. Impression of Georg Stiernhielm’s tree metaphor 164
13.1. Albert Schultens, 1730 172
13.2. Impression of Albert Schultens’s tripartite conceptual hierarchy 177
14.1. Johann Christoph Gatterer 184
16.1. Caspar Wyss’s Dialectologia sacra, 1650 209
18.1. August Schleicher’s (1863) family tree model 232
18.2. The linguistic continuum as a slanted line according to Schmidt 233
18.3. The linguistic continuum broken up according to Schmidt 233
xvi   
18.4. Hugo Schuchardt in 1922, picture by Hermann Urtel 234

18.5. Schuchardt’s (1928 [1868]: 165) conception of dialect groups visualized 236
20.1. Example of a typical bimodal distribution on a scale 260
21.1. Noam Chomsky in 1977 269
23.1. The Council of Europe’s headquarters in Strasbourg 291
List of Tables
4.1. Clement of Alexandria’s Greek compared to Hervet’s and Gessner’s

Latin versions 56
9.1. Transfer of social prejudices to the conceptual pair 134
11.1. Summary of the main early modern interpretations of the conceptual pair 148
14.1. Johann Christoph Gatterer’s four categories on the kinship continuum 189
Conventions
Non-English primary texts, when quoted, are systematically translated into English
in the main text. The original passages are given in the footnotes. Translations are
mine, unless otherwise indicated. Other relevant passages are occasionally cited in
the footnotes in their original form and without a translation. Ancient and medieval
Latin texts are quoted from the editions used by Brepols’s Latin databases (Library of
Latin Texts A and B as well as Monumenta Germaniae Historica), unless otherwise
indicated. I have regularized ancient Latin orthography by reserving <u> for the
vowel [u] and <v> for the semivowel [w]. Ancient Greek and Byzantine texts are
quoted from the editions used by the Thesaurus Linguae Graecae (TLG) database,
unless otherwise indicated. Greek words and short quotations have been transliter-
ated into the Roman alphabet in the main text, with the original Greek quoted
between round brackets. For recurring terms, the original Greek is given at the
word’s first appearance. Longer quotations are cited in the footnotes and provided
only in the original Greek alphabet. Titles of Greek and Latin texts cited in footnotes
have normally been taken over from the Brepols and TLG databases. All English
Bible quotes are cited according to the English Standard Version.
I have quoted early modern and modern texts as I have found them in the
original sources. I have resolved abbreviations between square brackets. Errors
and misprints have been marked with [sic]. Original bold font or unusual font, for
instance Gothic in an otherwise Latin-script text, have been converted to standard
Roman script. All italic fonts have been romanized. I have converted the super-
script e to the modern umlaut sign in early modern German quotations. When
pagination is lacking, I have used the signature markings to refer to the page
intended. The capital letters in titles of early modern works have been normalized
in the bibliography at the back. Abbreviations in the publisher’s names have not
been resolved. Names of Greek, Latin, and early modern authors have been
Anglicized whenever this is common in secondary literature. Otherwise, I have
opted for one of the most common forms. Life dates are provided in the main text
when a source author is first introduced but not in the footnotes. For living
persons, I offer the year of birth, if known. I refer to early modern dissertations
by mentioning the name of the supervising professor (praeses) as well as the
presenting student (respondens), unless I have good reasons to suppose that one of
them should be considered the sole author of the dissertation (on this issue see e.g.
Considine 2008b). Furthermore, to distinguish between concepts and terms, I rely
on italics to denote that the discussion pertains to a term. Finally, I use singular
they in order to avoid gendered language.
1
Introduction
When during World War II the linguist Max Weinreich (1894–1969) finished one
of his New York lectures on the Yiddish language and its social status, one of his
auditors, a young teacher from the Bronx, came up to the front of the room for a
follow-up discussion. He boldly addressed Weinreich: ‘What is the difference
between a dialect and a language?’—a question frequently asked of linguists in
general. As Weinreich tried to give him a satisfactory solution, the teacher
interrupted him and said: ‘That I know, but I will give you a better definition:
“A language is a dialect with an army and navy”.’ The identity of the auditor is still
up for discussion, but his remark was to have a rich and powerful resonance.
Weinreich was so impressed by the witty statement that he decided to mention
and discuss it at length in one of his publications, a paper, in Yiddish, entitled ‘The
YIVO and the problems of our time’, which he had prepared for the annual
conference of the YIVO, the Yiddish Scientific Institute, in 1945 (Weinreich
1945). The rest, as they say, is history. Today, as soon as the distinction between
the concepts language and dialect is raised in debate, someone will take refuge in
the quip recorded by Weinreich to relativize it and demonstrate that it is primarily
informed by language-external, sociopolitical circumstances. The question
whether one labels a speech form language or dialect depends, in other words,
on the history of its speech community, its social status, and its participation in
power or lack thereof (see Maxwell 2018).
In such debates, the distinction between language and dialect is, almost as
a rule, taken for granted. It is, in fact, usually perceived as a kind of ahistorical
given, an ever-present and obvious component of our metalinguistic apparatus
(cf. Kamusella 2016: 164). The conceptual pairing is, however, not as self-evident
as we tend to assume. Like all aspects of human culture, it has a history, which
I will sketch in this book. In this history, the early modern era, roughly 1500–1800,
was a critical stage. During this period, scholars with various interests and
backgrounds began to address, often at great length, questions about linguistic
diversity in general and dialectal variation in particular: why are there not only
diverse languages, but even differences in every language individually? Is there a
litmus test for determining whether a specific form of speech can be considered an
actual language or rather a dialect deriving from, and subsumed under, a lan-
guage? What criteria, causes, and circumstances can be invoked in this debate?
Despite the lively interest which early modern scholars took in these issues,
historians of linguistics have thus far made no attempt at exploring the origin and
Language or Dialect? The History of a Conceptual Pair. Raf Van Rooy, Oxford University Press (2020). © Raf Van Rooy.
DOI: 10.1093/oso/9780198845713.001.0001
2 
evolutions of the language/dialect pair in depth. Usually, it is situated in Greek

antiquity because of the origin of the term dialect (e.g. Kamusella 2015: 10–14;
2016: 172–5), but I will argue that this view is based on an unfounded assumption.
A rare exception is the all-round linguist Mario Alinei (1980: 11–27), who situated
the origin of the distinction in the Renaissance in an intriguing paper published
four decades ago. Alinei’s compatriot Mirko Tavoni (1986: 221), too, has placed
‘the birth of the modern concept [of dialect] through contact with the Greek
sources’ around 1500. In this book, I follow Alinei’s and Tavoni’s lead as I seek to
trace the emergence of the conceptual pair, a complex matter on the crossroads of
linguistics, intellectual history, and philology. My main aim is to present the
poorly known history of the language/dialect distinction to anyone interested,
inside and outside academia: linguists, historians, philologists, and non-specialists
with a general interest in history and language.
Even though the focus is on the early modern era, it is indispensable that I zoom
out. In order to fully grasp the history of the conceptual pair in Western linguistic
thought, I have to travel back in time to antiquity and the Greek world in
particular, with which early modern scholars developed a deep fascination.
There, they stumbled across an age-old situation of dialectal diversity that
reminded them of their native tongues; this encounter profoundly changed the
way they thought about variation within languages. Not only did they adopt the
term closely associated with Greek variation, diálektos (διάλεκτος), they also
adapted the concept that was intertwined with it to their own here-and-now. The
rediscovery of Greek diversity served as an eye-opener for humanists, many of
whom assumed that Latin, the language of science, education, administration, and
religion, was an invariable tongue that did not suffer from regional differences.
I do not only look back to antiquity, as this would result in a very incomplete
picture. Understanding the conceptual pairing, its different interpretations, and
the suspicions which many present-day scholars harbour about it requires an
investigation of modern developments, too. In this period, linguistics emerged as a
separate research discipline and was institutionalized. How did this affect the
language/dialect distinction? At the same time, the conceptual pair, until then
mainly restricted to educated circles, emerged as popular knowledge. In which
ways and contexts did this knowledge transfer occur, and why? By answering
these questions, it is possible to sketch the history of two intertwined concepts that
have been, and still are, frequently used by scholars and laypeople alike in a wide
range of different contexts.
My main focus is on the Western tradition of grammatical and linguistic
scholarship, with a slight West-Germanic tilt for five principal reasons. Firstly,
the conceptual pair seems to have been discussed most intensely by authors with a
West-Germanic background. Secondly, available scholarship is also concerned
with these authors. Thirdly, English texts are currently most easily accessible in
 3
this digital age. Fourthly, the focus is dictated in part also by the limitations of my
linguistic competence. Fifthly, a Germanic tilt also benefits the reader of this book,
obviously composed in English. The available source material is extremely varied,
especially for the periods before the modern age, when linguistics had not yet
emerged as an autonomous branch of study. It ranges from grammars, linguistic
handbooks, and treatises over lexica and journal articles to philological and
historiographical works of diverging nature. I focus on printed works for the
period from the Renaissance onwards, as these generally had a wider circulation
than manuscripts and therefore were in a position to achieve a greater impact.
In presenting the results of my analysis, I have tried to strike a balance between
the general and the specific. I have, on the one hand, taken a bird’s-eye view across
twenty-five centuries, tracing the journey of the language/dialect distinction from
its prehistory in antiquity to modern language studies. Charting continuities and
break-off points has been one of my main concerns. On the other hand, in order to
avoid doing injustice to the individuality of scholars and running the risk of losing
my grip on the actual sources, I have chosen to draw out telling case studies
throughout the book. As such, I have adopted an approach that is more or less in
line with David Armitage’s ‘history in ideas’,
a genre of intellectual history in which episodes of contestation over meaning

form the stepping-stones in a transtemporal narrative constructed over a span of
time extending over decades, if not centuries. The ‘ideas’ structuring this history
would not be hypostatised entities, making intermittent entries into the mundane
world from the idealism’s heavenly spheres, but rather focal points of arguments
shaped and debated episodically across time with a conscious—or at least a
provable—connection with both earlier and later instances of such struggles.
(Armitage 2012: 499)
This longue durée approach, which underlines the variability of concepts and
their context-bound constitution, has been further developed by Armitage in a
co-publication with Jo Guldi (Guldi and Armitage 2015). I have taken great care
to put the different episodes of the history of the conceptual pair in their correct
setting, an indispensable step in sound studies in intellectual history, as most
scholars agree following Quentin Skinner’s seminal 1969 paper ‘Meaning and
Understanding in the History of Ideas’. These contextualization efforts are all
the more crucial, since language was usually not studied in and of itself before
the nineteenth century. I take contexts as a plural noun, as they differ for each
author and each source text individually, and I follow Kristin Asdal and Ingunn
Moser’s suggestion that contexts should not be regarded as invariable givens
but rather as selective constructions of the historian’s own making (Asdal and
Moser 2012: 303).
4 
The main question which I aim to address in this book, then, is the following:
how and in which contexts have scholars thus far tried to distinguish a language
from a dialect? I have tried to answer this question in a more or less chronological
account of twenty-two brief chapters. This choice of presentation will, I hope,
make the complex history of the conceptual pair a palpable and well-structured
whole for the reader. Further coherence has been created by grouping the chapters
into five larger parts, which coincide with the main episodes of the story which
I am about to tell.
Part I, on Greco-Roman antiquity and the Middle Ages, is the shortest, and
with good reason, as I argue that the conceptual pair language/dialect was largely
absent from linguistic thought in these periods. This absence also explains why
I have called the first part ‘Prehistory’. Chapter 2 starts at the ultimate origin of the
term dialect, ancient Greece, unveiling how it was never customary in Greek
scholarship, both ancient and medieval, to contrast the term diálektos to a word
referring to language. In order to prove this, the chapter treats both passing
references to the Greek dialects and influential definitions of the Greek word. It
also frames definition attempts in their philological context. Finally, I briefly
discuss the Latin tradition up to about 1500, arguing that an obvious opposition
of dialect to language cannot be discovered there either.
Chapter 3 offers a detailed case study of the main exception to this general
tendency, the late medieval scholar and polyglot Roger Bacon, who opposed the
terms lingua and idioma in a way that prefigured the later language/dialect
distinction. Bacon was able to do so because of his exceptionally broad intellectual
and linguistic horizons; he was familiar with English, French, Latin, Greek, and
other tongues, as well as the regional variations within some of them. His mastery
of Greek in particular was unique in his times. Bacon’s linguistic outlook was a
central precondition and a triggering circumstance for his lingua/idioma distinc-
tion. He was in good company, since the renowned philosopher Thomas Aquinas
presupposed a similar metalinguistic contrast in his exegetical works; yet Thomas
was much less explicit about it. Chapter 3 finishes with a brief exposé of biblical
exegesis as an overlooked source for ancient and medieval ideas on regional
variation.
In Part II, I turn to the major argument of this book, endeavouring to reveal
the largescale emergence of the distinction between language and dialect in the
first half of the sixteenth century. It begins in the same way as Part I ended, with
a case study of a prominent scholar. Chapter 4 uncovers the way in which the
Swiss humanist Conrad Gessner, an important language scholar, bibliographer,
and zoologist, conceived of the Latin term dialectus in opposition to lingua.
Renaissance intellectuals were confronted with a major explosion of information,
also on the languages of the world, and Gessner was one of the first to try and
classify human speech in its great diversity. He did so in his Mithridates of 1555,
the first ever language catalogue, in which the term dialectus frequently appeared.
 5
The word served to bring more nuance into the relationships between speech
forms and is, unlike in antiquity and the Middle Ages, clearly taken to be a variety
of a language. In this regard, Gessner was inspired not only by ancient sources but
also by the works of his contemporaries. Indeed, unlike Roger Bacon, the Swiss
humanist was not an isolated pioneer, but rather the exponent of an early
sixteenth-century trend. The emergence of this trend is traced in Chapter 5
through its main symptoms. Most tellingly, the first decades of the Cinquecento
witnessed an increasing contrasting of the Latin terms dialectus and lingua, after
Greek diálektos had been definitively borrowed into Neo-Latin. Another symptom
was the creation of the concept of common language, which in its terminological
guise as lingua communis was often opposed to dialectus. In addition, dialectus
definitions were silently updated by means of determiners suggesting the hier-
archical subordination of the term vis-à-vis lingua. In the 1540s, a new Latin
phrase moreover appeared, ‘to differ only in dialect(s)’, which implied that related
dialects differed from each other only superficially, whereas distinct languages
exhibited substantial variation. This collocation enjoyed a rich career. Dialectus
and lingua were, however, not contrasted in a very clear fashion from the start, as
especially the former term was not yet delimited semantically. Instead, it was part
of an intricate conceptual web, including also concepts such as idiom and style.
Chapter 6 tries to frame the emergence of the language/dialect distinction in its
intellectual and historical context. The rediscovery of the Ancient Greek dialects
constituted, I argue, a major pivoting point, as well as the standardization projects
that were gathering steam in the early sixteenth century and the information
explosion enhanced by the commercialization of the printing press. Attention to
the standardization of the vernaculars was especially urgent in Protestant areas,
where a uniform language was needed to convey the Word of God to all classes of
society, while in Catholic Europe Latin remained the principal language of
religion. In the wake of standardization and categorization attempts, linguistic
diversity became countable, as boundaries became more fixed than before.
Paradoxically, humanists did not realize that they were introducing a new con-
ceptual pair, assuming, instead, that this had been a Greek achievement. Yet they
were, in fact, reading the language/dialect distinction into the ancient source texts.
As a result, the new contrast was a product of subconscious appropriation and
adaptation.
Part III, ‘Consolidation by elaboration’, describes how the new conceptual pair
became anchored in the metalinguistic apparatus of early modern scholars in the
century after 1550 and received various meanings, most of which are still in use
today. In Chapter 7, two interpretations with roots in Greek antiquity take centre
stage. First, I outline how the spatial conception of dialect established itself after
1550. More circumstantial evidence for the wide dissemination of this geograph-
ical interpretation is also briefly treated. Humanists moreover soon recognized the
universality of the phenomenon of regional linguistic variation, an intellectual
6 
achievement that has not yet received due acknowledgement. Even though the
geographical interpretation of the language/dialect distinction implied that
language covered a larger area than dialect, some early modern scholars believed
that good language had its seat only in a state’s capital. Secondly, I treat the
emergence of the early modern idea that dialect was particular to a tribe,
implying that language was a kind of ethnically overarching phenomenon.
There was, however, an unresolved tension with a competing view, associating
language with the nation in the political sense. Although widespread in the early
modern period, the spatial and ethnic conceptions of dialect as a variety of a
language were never used as diagnostic criteria to determine the language/
dialect status of a speech form. Chapter 8 treats two interpretations of the
conceptual pair which early modern scholars did consider useful in doing so:
the Aristotelian criterion and mutual intelligibility. I argue that in the seven-
teenth century an interpretation of the language/dialect distinction emerged
according to which related dialects showed only accidental differences, whereas
distinct languages varied in their substance. This analysis was grounded in two
traditional categories of Aristotelian ontology and implied a binary opposition
between substantial and accidental variation; numerous humanists realized,
however, that linguistic distance comes in degrees. In the remainder of
Chapter 8, I contend that the criterion of mutual intelligibility, often bracketed
together with the Aristotelian criterion, had its origin in the Renaissance. The
eccentric humanist Johannes Goropius Becanus played a key role in its emer-
gence and specified it further as immediate mutual intelligibility. This criterion,
too, was conceptualized in predominantly binary terms. Contradictorily,
numerous scholars complained about the lack of mutual intelligibility among
speakers of related dialects, pointing out the wider communicative reach of a
language.
From the more or less neutral early modern conceptions of the language/dialect
distinction in Chapters 7 and 8, I move to more subjective interpretations in
Chapter 9. From the late sixteenth century onwards, dialect was conceptualized as
an anomalous deviation from the analogical language under which it resided.
I suggest that this interpretation may have had its roots in Greek ideas on the
relationship between the Koine and the other dialects. The analogy/anomaly
conception was, however, principally grounded in early modern linguistic real-
ities, since the advancing standardization led to a stronger contrast between the
prescribed norm and everything deviating from it. Out of this normative inter-
pretation of the language/dialect distinction, the highly subjective idea that lan-
guage was superior to dialect developed almost naturally in the course of the
seventeenth century. Whereas theorizing on dialect had remained neutral in the
sixteenth century because of the close link it had with the esteemed Ancient Greek
dialects, the dialect concept was detached from them during the seventeenth
century. In addition, local elites gradually turned their back on their native dialects,
 7
embracing the upcoming standard languages. This social evolution likewise

enhanced the severe degradation of the dialect concept.
Most interpretations launched by early modern scholars were synchronic; they
concerned a snapshot in time only. Yet one take on the language/dialect distinc-
tion was diachronic, in that language was considered to generate different dialects.
The origin and diffusion of this historical conception is the subject of Chapter 10.
I argue that even though it had earlier precursors, it was only due to the influence
of Joseph Justus Scaliger’s work that this new interpretation became popular.
Already in the early seventeenth century, this conception of the language/dialect
distinction was moreover framed by Abraham Mylius within a cyclical process of
language change. The language-historical interpretation of the conceptual pair,
otherwise primarily understood in synchronic terms, soon prompted criticism,
voiced most fiercely by the seventeenth-century orientalist Johann Heinrich
Hottinger. I also comment briefly on the emergent idea that dialects preserved
archaic features, which apparently contradicted the diachronic conception of the
language/dialect distinction. Finally, I outline how the conceptual pair came to be
used as a handy discursive strategy for historical classifications of language,
especially in cases where evidence was lacking.
Chapter 11 gives an overview of how different definitions of the conceptual pair
emerged and became rooted in the thought of scholars. It does so, first and
foremost, by surveying the seven major interpretations of the language/dialect
distinction which originated roughly in the century between 1550 and 1650 and
the ways in which they interacted with each other. These conceptions were,
I argue, shaped by three main related circumstances: the Greek heritage, scholarly
interests, and sociolinguistic realities. I moreover develop the idea that in the
period 1550–1650 the conceptual pair underwent consolidation not only by
elaboration but also as an emancipation from the Greek heritage. This de-
Hellenization resulted in the degradation of dialect as secondary to language on
nearly all levels, even though a strictly political interpretation of the distinction
was only marginally present in the early modern period.
In Chapter 12, the last chapter of Part III, I present the third case study of the
book, the Swedish scholar Georg Stiernhielm. Driven by language-historical
interests, Stiernhielm defined the conceptual pair primarily in terms of substantial
versus accidental differences. This Aristotelian interpretation he made very expli-
cit, tying it to specific linguistic domains such as the lexicon and pronunciation.
He moreover invoked mutual intelligibility in his definitions. As he was concerned
in the first place with language history, his usage of the terms lingua and dialectus
was also steeped in the diachronic interpretation. More in the margin, the
analogy/anomaly opposition as well as geography likewise shaped his conception
of the distinction. The case of Stiernhielm, who probably did not know much
Greek, confirms the emancipatory evolution outlined in the previous chapter. It is,
finally, no coincidence that his interest in the conceptual pair surfaced around
8 
1650, immediately after he had met at the court of Queen Christina two erudite
scholars who had also reflected on the subject along similar lines: Christian Ravis
and Claude de Saumaise.
Stiernhielm’s extensive attention to the conceptual pair and his linking it to
linguistic features reveal a tendency towards systematization and rationalization.
Indeed, the Swedish philologist may well be regarded as a transitional figure,
heralding a new phase in the history of the language/dialect distinction, treated in
part IV. This next period coincides roughly with the years 1650–1800, the age of
rationalism and the Enlightenment, when the loose reflections of earlier times
were replaced by more structured discussions, and language scholars started to
involve more strictly linguistic elements in their treatments of the conceptual pair,
even though language-external factors were never far away. I first illustrate this
tendency towards systematization and rationalization by means of two short case
studies, before moving on to more general reflections. Chapter 13 outlines how the
eighteenth-century Dutch orientalist Albert Schultens repeatedly defined the term
dialectus in a highly systematic fashion. Schultens analysed the conceptual pair
principally in Aristotelian terms but tied it also to geographical factors and framed
it in a language-historical scheme. He moreover contrasted the analogy of lan-
guage to the anomaly of dialect. The Dutch orientalist extended the distinction so
as to include a third concept, that of degenerate offshoot, which, unlike a dialect,
did not preserve the core of the language intact. He also insisted on the linguistic
classes in which related dialects allegedly differed from each other. Schultens was a
key figure, as he put the conceptualization of dialect on the scholarly agenda, albeit
always as a matter of instrumental importance only, and triggered numerous
follow-up discussions among his pupils and readers.
Chapter 14 presents a short case study from an angle different from that of
philology, taking the historian Johann Christoph Gatterer’s ideas on linguistic
diversity as its object. It not only serves as another telling example of the tendency
towards systematization but also, and especially, represents a climax in
eighteenth-century attempts at rationalizing the conceptual pair. Proposing an
embryonic lexicostatistic method, Gatterer tried to find an objective way to use
linguistic data in writing an encompassing history of tribes and nations, in
particular their prehistory. Starting from a basic vocabulary set, Gatterer
attempted to quantify linguistic distance. In doing so, he divided the kinship
continuum into four sections: unrelated languages, related languages, dialects,
and closely related dialects. His innovative methodology, prefiguring modern
lexicostatistic approaches, had only limited success, however. Gatterer failed to
put it into practice, and the historian was criticized, rather ironically, for his
ahistorical method by the grammarian Johann Christoph Adelung.
In the last two chapters of Part IV, I adopt a broader perspective again.
Chapter 15 outlines the linguistic respects in which related dialects were believed
to vary, in opposition to distinct languages, during the early modern era. Initially,
 9
a Byzantine author was the main source of inspiration to describe dialect-level

differences, primarily in Greek handbooks. It was only after 1650 that the levels of
variation were treated in a more systematic fashion by non-Hellenists, too. The
focus of attention was on the ways in which related dialects varied. The differences
were, as most scholars agreed, superficial and mainly situated on the level of
pronunciation, letters, and the lexicon. There was, however, no linguistic domain
in which related dialects were claimed to never demonstrate variation. Overall, the
differences required to speak of distinct languages attracted less attention. Yet
many scholars agreed that substantial differences were needed, principally in the
roots of words. Sometimes, unusual linguistic criteria were put forward, for
instance by Johann Georg Wachter and Ferdinando Galiani in the eighteenth
century.
Chapter 16 discusses further evidence for the systematization and rationaliza-
tion of the language/dialect distinction in the period 1650–1800. On the one hand,
a kind of dialectological tradition emerged. The study of regional variation became
a subfield of philology, albeit never an autonomous one; occasionally, it already
was receiving the label of dialectologia, introduced in 1650. For the first time,
philologists presented dissertations on dialectal diversity that were no longer
exclusively focused on the Greek dialects. On the other hand, scholars adopted
more rational attitudes towards the conceptual pair. Some chose to supplement
the binary contrast with new concepts. Others advocated to distinguish more
clearly among different interpretations of the language/dialect distinction.
Confusion persisted, however, throughout the early modern period. The first
vocal sceptic of the conceptual pair was Friedrich Carl Fulda, who made painfully
clear how arbitrary the distinction actually was.
In Part V, I explore the vicissitudes of the conceptual pair in mainstream modern
linguistics, with specific attention to their relationship to earlier ideas, both in
terms of continuity and breaking points. Chapter 17 surveys the nineteenth-
century fate of the language/dialect distinction, during and after the establish-
ment of linguistics as an autonomous field of study. As at this stage the study of
language roughly coincided with historical-comparative grammar, it is no sur-
prise that the language-historical interpretation prevailed. In the 1870s, the
spatial conception was foregrounded when dialect geography emerged. Other
criteria, too, lived on, especially mutual intelligibility and the Aristotelian
interpretation. In the first half of the nineteenth century, only few scholars
bothered to discuss the conceptual pair at length, with the exception of the
forgotten comparative linguist and Hellenist Albert Giese. In the 1860s, suspi-
cion about the validity of the language/dialect contrast grew, perceptible in the
work of August Schleicher and the neogrammarians, but not to such an extent
that it was extensively questioned. Perhaps linguists were afraid that this debate
would mean the premature end of their young discipline. By the end of the
nineteenth century, many prominent linguists had realized the arbitrariness of
10 
the language/dialect distinction, and linguists were forcibly confronted with its
problematic nature from within, as Chapter 18 argues. This last case study of the
book outlines the ideas of the headstrong German linguist Hugo Schuchardt,
who dared to face the full consequences of critically reflecting on the conceptual
pair. Considering the nature of linguistic diversity, Schuchardt noticed that
there were no such strict divisions as Schleicher’s family tree model presup-
posed. Instead, he proposed a wave model, thus prefiguring Johannes Schmidt’s
image. Language and dialect were merely fictional abstractions from actual
linguistic facts and for this reason useless. Schuchardt was not alone in his
suspicion. Some of his colleagues likewise acknowledged the abstract nature of
the language/dialect distinction, but preferred to keep on using it for practical
reasons, whereas others like Jules Gilliéron dismissed its validity and concen-
trated on linguistic features instead, realizing that dialect boundaries were
arbitrary.
Chapter 19 surveys the fate of the language/dialect distinction in structuralism.
Ferdinand de Saussure’s conception of it is revealed to have been fairly traditional.
In Saussure’s wake, mainstream structural linguists usually focused on homoge-
neous language systems, the langue, rather than the parole, with scant attention to
the conceptual pair. In the 1950s, a dialectological turn occurred. The year 1954 in
particular was a breaking point, when three structuralist papers devoted to the
concept of dialect appeared. Uriel Weinreich suggested the concept of diasystem
to capture variation within one language. André Martinet, in turn, tried to
redefine dialect scientifically by excluding sociopolitical factors. Václav Polák,
finally, argued that substantial morphosyntactic variation was required to speak
of distinct languages. Phonological and lexical differences resulted in dialects only.
Structuralist discussions of the language/dialect pair remained uncoordinated,
however, and had relatively limited impact on subsequent debates, except for
Weinreich’s diasystem concept.
The brief Chapter 20 treats the success of the criterion of mutual intelligibility
since the 1950s, when American linguists interested in Amerindian tongues
started to actively test this feature. Pioneers were Carl Voegelin and Zellig
Harris, who suggested four methods of answering language/dialect questions,
including mutual intelligibility testing. Even though scholars immediately faced
numerous problems, the method enjoyed considerable success and is the primary
criterion used by language catalogues such as Ethnologue and Glottolog. The
criterion was criticized by, among others, Frederick Agard, who proposed nine
postulates for determining language/dialect status. Others followed Morris
Swadesh’s lexicostatistic lead and tried to quantify the distance between speech
forms. According to one of the most recent representatives in this strand, linguis-
tic distance is bimodally distributed, and the language/dialect distinction is, by
consequence, backed by the majority of linguistic evidence. Rather artificially, a
strict cut-off point is proposed between the language and dialect poles.
 11
Chapter 21 confronts two opposite perspectives on language and the language/

dialect pair with roots in the 1950s and 1960s. On the one hand, it treats
generative approaches to linguistic diversity. In general, generative linguists
have assumed that dialect-level variation is produced by minor differences in
rules, parameters, or constraints and their ordering or ranking, depending on the
generative framework which they follow, whereas distinct languages are charac-
terized by major divergences in the same. On the other hand, sociolinguists have
focused on linguistic variables rather than systems. As they correlate linguistic
phenomena with language-external attributes, their conceptions of the language/
dialect distinction tend to be rather hybrid, being shaped by linguistic as well as
sociopolitical parameters. This externalist approach has been fiercely criticized by
Noam Chomsky, the founding father of generativism. Other linguists have
adopted more constructive attitudes, either by supplementing the language/dialect
distinction or by supplanting it with an entirely new conception of language.
Chapter 22 develops the idea that even though many linguists today tend to
avoid the term dialect, it has been a major resource for new terminology, especially
through -lect derivations. The recent terms doculect and languoid are particularly
interesting as they are derived from (dia)lect and language, respectively, and are
part of an alternative conception of language. Terminology linked to the language/
dialect distinction is still widely used and has been incorporated into conceptual
constellations tailored to specific national contexts. It is moreover striking how
noncommittal recent definitions of dialect are, as if linguists want to keep a safe
distance from this tarnished term. The final section draws the balance of the
modern love-hate relationship with the conceptual pair. Five main attitudes can be
distinguished, ranging from tacit preservation through explicit redefinition and
supplementation to a quest for alternatives and downright dismissal. The
unorganized nature of the debate on the subject is remarkable, as is the persistence
of centuries-old interpretations.
Even though the book’s focus is on scholarly conceptions of the language/
dialect pair, first within grammar and philology and later within linguistics, it is
indispensable to dwell briefly on popular ideas, too, since they have interfered and
interacted with the views of linguists in various intriguing ways. Chapter 23
explores the modern popularization of the conceptual distinction, gathering
steam in the last decades of the nineteenth century, when the standard language
ideologies of nation-states spread to nearly all classes of society, primarily through
mass education. This popularization involved a sharp divide between standard
language and dialect, which became widely known among the general public. In
the process, the conceptual pair was excessively politicized and even politically
activated, as demonstrated by some notorious episodes in recent history, including
the Oakland Ebonics controversy. The Internet has, moreover, provided laypeople
with a forum in which to debate the language/dialect distinction and its applica-
tions, which tend to have a strong political colour. It remains to be seen whether
12 
and how linguists and others, both laypeople and academics, have exchanged
ideas on the subject. I tentatively suggest that there indeed has been a conceptual
cross-fertilization, a phenomenon requiring further investigation.
Chapter 24, finally, surveys by way of conclusion the book’s main arguments.
These include especially the emergence of the modern language/dialect distinction
during the early sixteenth century and the subsequent formulation of its main
interpretations. Above all, however, this chapter emphasizes that the conceptual
pair unmistakably has a history, for too long neglected, and that it is not a self-
evident given which has always been there. Having established the historicity of
the language/dialect distinction, I field the question of whether it has a future, to
which I try to offer an answer, both tentative and brief, from my perspective as a
historian of language studies. On the one hand, I suggest that a reconceptualiza-
tion of the distinction can be a viable option. On the other hand, the fact that the
conceptual pair has become common knowledge in modern times gives linguists,
I believe, not only the opportunity but also, and especially, the responsibility to
take on a more prominent societal role in language/dialect disputes.
I
P R E H I S T O R Y , 5 0 0   – 1500
2
A dive into the prehistory
of the conceptual pair
As the Roman emperor Tiberius (42 – 37) was very fond of Greek rhetoric,
he allowed several Greeks to be part of his entourage and live with him. If we are to
believe the historian Suetonius (c.  70–after  128), however, they were subject
to his capriciousness as much as anyone else. One of Tiberius’s favourite Greeks,
going by the name Zeno, for instance, suffered a tragic fate after displeasing the
emperor. The Greek had uttered some phrases in an affected fashion, which
bothered Tiberius, who asked: ‘What is that so very annoying dialect?’ Zeno
answered that it was Doric, which caused the emperor to promptly banish him
to the Greek island of Cinaria, present-day Kinaros in the Dodecanese. Zeno had
apparently reminded Tiberius of the miserable period in his life when he lived in
retirement on the Doric island of Rhodes, before he was crowned emperor.¹ For
this reason, he wanted him gone.
Suetonius inserted this episode into his biography of Tiberius as he wanted to
demonstrate the ruthless character of the Roman emperor. Indirectly, it indicates
something else, too. It was widely known in the intellectual circles of antiquity,
even in the Latin-speaking western sphere of the Roman Empire, that there were
different Greek dialects. Indeed, they absorbed nearly all attention turned to the
topic of variation within a language before about 1500, since they were the only
dialects relevant for literary and philological study. Greek literature was com-
posed not in one uniform language but in a range of different dialects (see e.g.
Colvin 2010). Homer’s masterpieces were sung in a dialectally mixed
Kunstsprache. The great orators from ancient Athens drew up their speeches in
Attic, the classical form of Greek par excellence. Pastoral poetry was marked by
Doric dialect, whereas Sappho chose her native Aeolic speech in her touching
lyrical songs. Herodotus, the proclaimed father of historiography, wrote his
histories in Ionic.
¹ Suetonius, De vita Caesarum, Tiberius 56.1: ‘Nihilo lenior in convictores Graeculos, quibus vel
maxime adquiescebat, Xenonem quendam exquisitius sermocinantem cum interrogasset, quaenam illa
tam molesta dialectos esset, et ille respondisset Doridem, relegavit Cinariam, existimans exprobratum
sibi veterem secessum, quod Dorice Rhodii loquantur’.
DOI: 10.1093/oso/9780198845713.001.0001
16      
2.1 The Greek dialects between anecdotes and definitions
Suetonius’s anecdote about Tiberius and Zeno is a fairly banal instance of an

ancient author referring to the Greek dialects.² As such, it represents a rather low
level of awareness of dialectal variation. Several centuries before Suetonius, the
Greek historian Herodotus (c. 485–424 ) was more conscious of Greek diver-
sification, as he was in a position to distinguish four different varieties of Ionic
Greek in Asia Minor, part of present-day Turkey. He refers to them as trópous
tésseras paragōgéōn, ‘four manners of deviations,’, and kharaktêres glōś sēs tésseres,
‘four distinctions of tongue’.² Herodotus’s wording indicates that he could not yet
rely on an established conceptual apparatus and a corresponding metalanguage
to talk about dialectal diversity. Instead, he had to resort to words lacking an
obvious semantic link with language at that time, such as kharaktē ŕ (χαρακτήρ),
‘character(istic); distinctive mark; stamp’, trópos (τρόπος), ‘way; manner’, and
paragōgē ́ (παραγωγή), which for Herodotus apparently meant something like
‘deviation’, ‘twisting’, or ‘seduction’. Paragōgē ́ did later become a metalinguistic
term meaning ‘derived form’ and ‘inflection’, whereas the root kharak- featured
prominently in later Greek definitions of the term diálektos.³
Some generations later, the Athenian general and notoriously difficult histori-
ographer Thucydides (second half of the fifth century ) went a little further still
by trying to make sense of the different Greek dialects and to characterize their
interrelationships. He reported, for instance, on a case of dialect mixture on Sicily
(Historiae 6.5.1). The inhabitants of the city of Himera spoke, Thucydides
claimed, a variety that occupied a middle ground between Chalcidian Ionic and
Doric, an assertion for which there is, by the way, no historical evidence (Vassallo
2005: 89). The Athenian historian also mentioned that the Greek spoken by the
Aetolians was not understood by other Greeks (Historiae 3.94.5), thus apparently
proving that not all varieties of Greek were mutually intelligible.
Another type of awareness of dialectal variation came to the fore in its dramatic
and literary activation. The brilliant playwright Aristophanes (c. 450–c. 385 ),
for instance, put specific dialects to comic use at many different occasions, as
proved by the masterful analysis of Stephen Colvin (1999). Connected to this was
the assumption that a dialect conveyed a specific impression to a listener or reader,
which surfaced in other genres, too. For example, the great philosopher Plato
(428/7–348/7 ) apparently experienced the Lacedaemonians, Greeks of Doric
descent, as brakhúlogos (βραχύλογος), ‘short in speech’ (Leges 641e). A feature
frequently attributed to Dorians in general was their alleged ‘broadness’ in speaking
² Historiae 1.142: ‘τρόπους τέσσερας παραγωγέων’ & ‘χαρακτῆρες γλώσσης τέσσερες.’ Cf. Van Rooy
(2016d: 247–8).
³ For paragōgē ́ see Liddell & Scott (1996: s.v. παραγωγή); Dickey (2007: 251). For the kharak-root
see Van Rooy (2016b: 60–1).
.    . . . 17
(e.g. Theocritus, Idyllia 15.87–8). These views are among the earliest explicit
formulations of the impression which a dialect conveys or, in modern linguistic
terminology, of dialect attitudes and perceptions (cf. Edwards 2009: 82–97).
The rhetorician and lexicographer Julius Pollux (second half of the second
century ) elaborated in more general terms on the variability of language by
pointing out that differences in tongue existed from city to city (Onomasticon
2.110). Before him, the geographer Strabo (c. 62 –c.  24) had made a similar
suggestion (Geographica 8.1.2). Ancient Greek authors undertook no real
attempts at explaining variation in their native language, even though Diogenes
Laertius (mid-third century ), author of a work on the lives and thought of
various philosophers, suggested colonization as a force behind linguistic change
(Vitae philosophorum 1.51).
In the past few paragraphs, I have covered the spectrum from vague awareness
of dialectal variation to a preliminary search for the causes of the phenomenon.
With Diogenes Laertius, I can even go one step further and reach a still higher
degree of abstraction. For in his work we find one of the earliest definitions of the
key term diálektos, which he quoted from a now lost work by Diogenes of Babylon
(c. 240–150 ). This Stoic philosopher must have explained the word as follows:
Dialect is speech stamped tribally and in Greek fashion, or speech of a certain

region, that is, having a certain quality according to a dialect, as thálatta
[θάλαττα, ‘sea’] in the Attic and hēmérē [ἡμέρη, ‘day’] in the Ionic.⁴
Diálektos was conceived by Diogenes as a technical term to designate features of

the Greek language that bore the mark of a certain tribe or region. The philoso-
pher expressed this by means of the kharak-root, as Herodotus had done several
centuries earlier. The verb kharássō (χαράσσω) meant, among other things, ‘to
carve; to engrave; to stamp’, specifically with reference to coins. By using this verb,
Diogenes probably wanted to emphasize the distinctive character of a diálektos,
and how a tribe and its members could leave their own mark on features of the
Greek language. He provided specific examples, too. Thálatta, as opposed to
thálassa with double sigma, was a Greek word bearing the stamp of the Attic
tribe and the region of Attica, whereas hēmérē, as opposed to hēmérā (ἡμέρα), was
typical of the Ionic dialect.
Even though Diogenes of Babylon linked diálektos with regional variation, he
did not presuppose a clear-cut distinction between dialect and language. Instead,
the term was understood as léxis (λέξις), which Diogenes explained very generally
⁴ Diogenes Laertius, Vitae philosophorum 7.56: ‘Διάλεκτος δέ ἐστι λέξις κεχαραγμένη ἐθνικῶς τε καὶ
Ἑλληνικῶς, ἢ λέξις ποταπή, τουτέστι ποιὰ κατὰ διάλεκτον, οἷον κατὰ μὲν τὴν Ἀτθίδα θάλαττα, κατὰ δὲ
τὴν Ἰάδα ἡμέρη’ (translation adapted from Van Rooy 2016d: 250).
18      
as ‘speech that is expressible by means of letters’, phōnē ̀ eggrámmatos (ϕωνὴ

ἐγγράμματος) in Greek (Diogenes Laertius, Vitae philosophorum 7.56). Accounts
such as Diogenes’s, which testify to a larger degree of abstraction than shown by
previous observations, will constitute the central thread running through this
chapter, as they are indispensable for an adequate understanding of numerous
aspects of later theorizing on the language/dialect distinction.
2.2 Diálektos, a variety of interpretations
Was Diogenes of Babylon alone in trying to define diálektos? Or was he part of a

tradition? The scant evidence available suggests that the latter is true. Indeed,
numerous Greek scholars tried to define diálektos as a workable metalinguistic
term, albeit in different meanings (Van Rooy 2016b, 2016d). Aristotle’s (384–322 )
early interpretation of diálektos as ‘articulation of the voice by means of the
tongue’ is miles away from later definitions focused on the literary varieties of
Greek.⁵ Yet even in works devoted to the Ancient Greek dialects, there was
considerable disagreement.
The term diálektos had an important generic meaning referring to language,
‘manner of speaking; means of communication’, deriving from the Greek middle
verb dialégomai (διαλέγομαι), ‘to converse with’. With this generic meaning, the
widespread definition idíōma glōssēś or glōt́ tēs (ἰδίωμα γλώσσης/γλώττης), found
in several philological and grammatical works on Greek, shows affinity. In my
view, the definition must have signified ‘particularity of tongue’ rather than
‘particular form of a language’ (pace Consani 2000: 612–13). The difference
between the two interpretations lies in the presence or absence of conceptual
subsumption. I will make use of the terms subsumption and to subsume in this
book to refer to a key development in the history of the conceptual pair language/
dialect: the emergence of the idea that there is a level of linguistic entities, dialects,
which fall under the hierarchically higher category of languages. In other words, at
a certain point in history, the concept of dialect was subsumed under that of
language, an issue which I will address at length in Part II.
According to my interpretation, then, the widespread Greek definition idíōma
́
glōssēs meant ‘particularity of tongue’, in which case diálektos was not subsumed
under a superordinate language concept. Instead, it was viewed as a language
particularity in a very general sense, as an ensemble of language features that did
not have a wide application, but were restricted to a certain context of usage.
Doric, for instance, was a diálektos because it had long alphas where one would
normally expect etas, because it was particular in contracting the vowel sequence
⁵ Historia animalium 535a: ‘διάλεκτος δ’ ἡ τῆς ϕωνῆς ἐστι τῇ γλώττῃ διάρθρωσις’.

. ́,     19
epsilon-omicron <εο> as epsilon-upsilon <ευ>, etc. Indeed, nowhere in extant

Greek texts are there any indications that the Greeks understood the definition
́
idíōma glōssēs as referring to a ‘particular variety of the Greek language’, and, by
consequence, it is impossible to attribute a subsumption of dialect under language
to Ancient Greek and Byzantine thought. Nonetheless, the definition idíōma
́
glōssēs was, just like the generic meaning of ‘manner of speaking’, intrinsically
linked to linguistic diversity. After all, speaking about a ‘manner of speaking’ or a
‘particularity of tongue’ presupposed the existence of other ‘manners of speaking’
and ‘particularities of tongue’. The definition idíōma glōś sēs is, for example, found
in the influential works of John the Grammarian, to be identified with the early
Byzantine philologist and theologian John Philoponus (?c.  490–575), and the
Byzantine polymath Gregory of Corinth (eleventh / twelfth centuries), also known
as Gregorius Pardus.
Several ancient and Byzantine scholars further specified the semantic core of
the word diálektos in their definitions of the term. What parameters did they
invoke? Greek thinkers laid down three principal specifications. First and
foremost, a diálektos was interpreted as speech particular to a certain Greek
éthnos (ἔθνος), ‘tribe’. This parameter already appeared in the early definition
by Diogenes of Babylon, quoted in the previous section. We also encounter
it in the work of, among others, the Early Christian author Clement of
Alexandria (c.  140/50–before 215/21) and Gregory of Corinth.⁶ These
scholars consistently relied on the kharak-root to make clear that a diálektos
always bore the ‘stamp’ of a Greek tribe. Their frequent recurrence to this root
seems to point to the existence of a continuous tradition, starting with
Diogenes at the latest and having a precursor in Herodotus’s usage of the
term kharaktē ŕ . Clement of Alexandria is an exception in that he did not
restrict éthnos to the Greek tribes but extrapolated it to Hebrew, the language
of the Old Testament, too (Stromata 6.15.129.2). This broad use suggests once
again that diálektos did not figure in a hierarchic constellation in which it was
subsumed under a term like glôssa or glôtta (γλῶσσα/γλῶττα), meaning ‘lan-
guage’. Instead, it could refer to both Hebrew and Ionic, speech forms we
would be tempted to call from our modern perspective language and dialect,
respectively.
Second, a diálektos was speech particular to a certain tópos (τόπος), ‘place’. This
specification was indisputably present in ancient definitions but seems to have lost
its importance in medieval times. It was, for instance, absent from important
Byzantine reference works such as the Suda. What is more, it did not figure in
John the Grammarian’s and Gregory of Corinth’s definitions, in which the Greek
word tópos, ‘place’, came to be strangely replaced by túpos (τύπος), ‘type, form’.
⁶ Clement of Alexandria, Stromata 1.21.142.3 & 6.15.129.2; Gregory of Corinth, De dialectis 1.1.
20      
This change was probably no coincidence and might reflect an inability to tie the
Greek dialects to geographical locations, since the dialects had been extinct for
quite some time when these authors were writing down, or rather incorrectly
copying, their definitions (cf. Dickey 2007: 75; Tribulato 2014: 458–9). Still, in
the early Byzantine period, John the Grammarian did regard the geographical
dispersion of Greek tribes as a force behind dialectal diversification (Van Rooy
2016d: 264–5).
A diálektos, then, was for many Greek scholars ‘speech of a particular form, tied
to a specific Greek tribe and place’. Oddly enough, the Greek Koine was usually
also classified as one of the main dialects, even though this variety could not
be linked to a distinct tribe or region. For the Koine, short for hē koinē ̀ diálektos
(ἡ κοινὴ διάλεκτος), ‘the common way of speaking’, was a supraregional form of
Greek which in the wake of Alexander the Great’s conquests spread across the
eastern Mediterranean and beyond, where it served as a lingua franca. This
incompatibility remained, however, unresolved in Greek thought, even though
there are traces of a debate on the exact position of the Koine vis-à-vis the other
dialects (Van Rooy 2016d: 257–8).
Thirdly, Greek scholars seem to have regarded a diálektos as written speech
particular to an author or a group of authors. They did not take the Greek spoken
by the man in the street into consideration, as it was all too susceptible to change
and probably difficult to write down in all of its phonic particularities. On the
contrary, grammarians concentrated on speech that was expressible by means
of letters—Diogenes of Babylon’s phōnē ̀ eggrámmatos—and on the canonical
authors, who happened to compose in different forms of Greek but at the same
time avoided features of too regional a colour. It was this coincidence of Greek
literary history that made the Greek dialects relevant to philologists (e.g. Mickey
1981; Morpurgo Davies 1987: 10–11). The literary status of the Greek dialects
entailed a close association of the word diálektos with written codification and
made it also applicable to other languages with a well-established written trad-
ition, such as Latin, Egyptian, and Hebrew.
The playwright Aristophanes and the Byzantine archbishop Michael Choniates
(c. 1138–c. 1222) were exceptional in that they used the term diálektos to refer to
varieties spoken by lower social classes (Van Rooy 2016d: 258–9, 266–7).
Consequently, the word clearly did not have a negative connotation, contrary to
many later uses of the English term dialect and its equivalents in other languages.
The lexicographer of rare words Hesychius (?fifth/sixth centuries ), for
instance, needed the adjective parátonos (παράτονος), ‘ill-sounding’, to add a
pejorative meaning to his explanation of barbarismós (βαρβαρισμός), ‘barbarism’,
as ‘parátonos diálektos’ (‘παράτονος διάλεκτος’ at Lexicon .210). As a result, the
word was often absent from contexts in which we would today intuitively expect it
to be present. The Greek geographer Strabo, for instance, described variation in
the tongues of Gaul by stating that the Belgians and the Celts ‘do not all speak the
.    21
same language, but some of them slightly diverge in their tongues’.⁷ The fact that
Strabo was discussing here non-Greek, unwritten tongues might have precluded
the use of the word diálektos. Also, the term was not yet interpreted as a variety of
a language differing only superficially from other varieties of the same language, as
it later was.
2.3 Philology as stimulus
Greek scholars proposed various definitions of diálektos, but why did they care
to define the term at all? The Greek dialects were relevant only in so far as they
were the forms in which the great literary works of Homer and others were
composed. It is consequently impossible to identify professional Greek dialect-
ologists. No one devoted his efforts to studying the Greek dialects in and of
themselves. Furthermore, the word dialectology, although entirely composed of
Greek lexemes, was only coined in the mid-seventeenth century (see Chapter 16,
Section 16.1). Still, it remains a fact that treatises devoted to the particularities of
the literary dialects, mainly Attic, Ionic, Doric, and Aeolic, and sometimes also
including the Koine, were composed at an early stage. In these works, every
dialect was discussed separately, and the letter mutations occurring in each of
them took centre stage. For instance, to use an example already cited, most
treatises pointed out that Attic had double tau where other dialects had double
sigma, as in thálassa versus thálatta, ‘sea’. These changes were conceived as
‘modifications of the word’, páthē tês léxeōs (πάθη τῆς λέξεως) in Greek, an
approach encompassing several different letter change operations and usually
called ‘pathology’ by modern scholars (e.g. Wackernagel 1876; Siebenborn 1976:
150; Ax 1987; Lallot 1995). Early works in this philological tradition by the
grammarians Tryphon (second half of the first century ), Apollonius Dyscolus
(first half of the second century), and others are lost to the ages. The first entirely
extant writing in this tradition might be pseudo-Plutarch’s Life of Homer. In a
part of this work, the Greek dialects as they appear in Homer’s epic poems are
briefly described. This section came to be extracted from the treatise, most likely
in Renaissance Italy, and often received the title On the dialects in Homer. The
treatise may date from the Roman period, but this issue is complicated by the
fact that it has received later additions (Van Rooy 2018c). Several other but
clearly related treatises survive, most of which have been collected by Gottfried
Heinrich Schäfer (1764–1840) in the early nineteenth century (Schäfer 1811).
One treatise, John the Grammarian’s Tekhniká (Τεχνικά), which can be trans-
lated roughly as Grammatical issues, deserves special attention here. Probably
⁷ Geographica 4.1.1: ‘ὁμογλώττους δ᾽ οὐ πάντας, ἀλλ᾽ ἐνίους μικρὸν παραλλάττοντας ταῖς γλώτταις’.
22      
based on Alexandrine sources and extant in two different redactions, the work
exhibits two features of interest as it stands.⁸ On the one hand, the second
redaction elaborates on the three linguistic levels on which the Greek dialects
allegedly exhibited differences: of entire words, of parts of words, and of word
accidents, relating to matters of accent and other diacritical marks. The focus was,
however, clearly on the second category of variation, which primarily covered
letter mutations. John the Grammarian’s account influenced many early modern
discussions of the levels on which related dialects varied, as I will argue in a later
chapter (Chapter 15, Section 15.1). On the other hand, the link with philology and
the study of literature surfaces very clearly when at the very end of the first
redaction, individual authors—exclusively of pagan stock—are connected to spe-
cific dialects.⁹ Such associations could be incorrect, as is clear from the fact that the
Greek poet Pindar (?522/18–after 446 ) is wrongly said to have written in the
Koine, even though the Koine did not yet exist at the time of his writing and his
poems have a primarily Doric look (Morpurgo Davies 1987: 18).
Philological motives also informed classifications of the Greek dialects (Van
Rooy 2016a; 2020a). If Attic and Ionic are today believed to constitute one dialect
branch of Greek, ancient and Byzantine scholars separated them based on their
distinct literary usages. The position of the Koine vis-à-vis the four literary
dialects, moreover, remained a contested issue throughout antiquity and the
Byzantine era, as I have mentioned earlier. This struggle with the status of the
Koine inhibited the wide acceptance of the idea that the Koine was the normative
variety from which the dialects were deviations. As a matter of fact, a standard
language concept was absent from Greek theorizing, which does not mean,
however, that there did not exist a notion of correct usage among Greek gram-
marians. There was a tradition of normative linguistic thought separating correct
from incorrect forms of Greek (Versteegh 1986; Dickey 2007: 235). The position
of the dialects in this correct/incorrect dichotomy nevertheless remained some-
what unclear. Overall, the Greek linguistic ideal of Hellēnismós (Ἑλληνισμός),
‘Greekness’, usually encompassed the canonical literary dialects other than the
Koine, too, as James Clackson (2015) has pointed out.
2.4 The Latin West
What about the Latin West? How did ancient Roman and Western medieval
scholars approach regional linguistic variation? There, the linguistic context was
of a very different nature than that in Greece. There was only one language of
⁸ On the work’s sources see Bolognesi (1953: 101–2); Hainsworth (1967: 63–4).
⁹ See John the Grammarian in Manutius et al. (1496: 236-). I refer to the first Renaissance edition
because, still, no modern critical edition of this important text is available.
.    23
literature and administration: Latin, an Italic tongue that had succeeded in largely
wiping out the great linguistic diversity in the western half of the Roman Empire.
This situation gave rise to a stricter dichotomy between correct and incorrect
language. Ancient authors nonetheless had an eye for regional differences within
Latin. To give only one early example, I can mention a passage in one of Plautus’s
(c. 250 –after 191 ) comedies. In his Trinummus, the playwright mocked at a
phraseological particularity of the city of Praeneste in Latium, present-day
Palestrina, some thirty-five kilometres east of Rome. When the slave Stasimus
tried to communicate to his master Callicles that his son had betrothed his
daughter to a certain Lysiteles, his baffled master asked him when this had
happened. Stasimus then replied: ‘just now (tam modo), as the Praenestine
says’.¹⁰ The use of tam modo instead of modo bore, in other words, the mark of
Praeneste, which Plautus put to comic use.
Contrary to what present-day readers might expect, Roman authors did not
intuitively compare the Greek literary varieties with regional variation in their
native language but linked the Greek dialects with different linguistic registers in
Latin instead. The Roman rhetorician Quintilian (c.  35–100) provided a
powerful testimony to this association, when discussing so-called Sardism. This
fallacy was named after the city of Sardis, situated in modern-day Turkey, which
in antiquity was known for its dialectally mixed population. Quintilian explained
it as follows:
Also, Sardism is the name of a certain speech mixed from a diverging range of
tongues, if, for instance, you would confound Doric, Ionic, or even Aeolic words
with Attic ones. But we have a similar vice, if someone mixes lofty with lowly
words, old with new ones, and poetic with vulgar ones—that is indeed such a
monstrosity, as Horace writes in the first part of his book on the art of poetry: “if
a painter would want to join a horse’s neck to a human head”—and would place
other things of different natures under it.¹¹
Whereas the Greek example referred to regional varieties that had been elevated to
literary status, the Latin situation did not pertain to regional linguistic diversity at
all but to different stylistic registers, some of which were linked to time-bound and
class-based language variation. It can be noted here that the term diálektos did not
feature in Quintilian’s comparison, but the orator did use the word when talking
¹⁰ Trinummus 609: ‘ “tam modo” inquit Praenestinus’.

¹¹ Institutio oratoria 8.3.59: ‘Σαρδισμός quoque appellatur quaedam mixta ex varia ratione lin-
guarum oratio, ut si Atticis Dorica, Ionica, Aeolica etiam dicta confundas. Cui simile vitium est apud
nos, si quis sublimia humilibus, vetera novis, poetica vulgaribus misceat—id enim tale monstrum, quale
Horatius in prima parte libri de arte poetica fingit: Humano capiti cervicem pictor equinam iungere si
velit—et cetera ex diversis naturis subiciat’. On the ancient concept of Sardism see Gitner (2018).
24      
about the Greek dialects on two different occasions (Institutio oratoria 1.5.29 and
9.4.18). Elsewhere (11.2.50), he referred to the five Greek dialects as Graeci
sermonis differentiae, ‘differences of Greek speech’. Quintilian’s choice of words
was part of a broader tendency. The Greek word diálektos as well as its Latinized
counterpart dialectus were utterly rare in Latin texts before 1500, and where they
did appear, they almost always denoted a variety of the Greek language (Van Rooy
2019). What is more, no ancient or medieval Latin author contrasted this term
with a word meaning ‘language’.
The case of Quintilian suggests that at least some educated ancient Romans
were familiar with variation in both Latin and Greek. As I will argue in later
chapters (Chapter 3, Section 3.3; Chapter 4, Section 4.1), an awareness of different
languages and their internal variation seems to have constituted an important
trigger for developing a language/dialect-like distinction. It was, however, by no
means a sufficient condition, as ancient Roman authors did not develop a con-
ceptual opposition resembling such a distinction (pace Kamusella 2015: 11–13;
2016: 172–5). Why did this outlook not act as a trigger in the case of the Romans?
The answer probably lies in the fact that the Greek dialects were perceived as
resembling Latin literary registers rather than Latin regional variation, as
Quintilian’s comments on Sardism indicate. There was, moreover, a general lack
of interest in the regional and social diversification of Latin, with the bulk of
attention being directed to so-called good Latin. What is more, Roman and early
medieval authors probably did not view Latin variation as constituting clearly
demarcated dialects each with a distinct set of properties, as was the case with the
Greek dialects. As a result, they provided only very few generalizations with regard
to Latin variation and refrained from offering strict definitions of the metalin-
guistic terms they used in this context. One might nevertheless note here that in
the Greek world, variation other than the one observed among the canonical
literary dialects received still less attention than variation in the Latin sphere
(Kramer 1989; Müller 2001: 279 n.17).
The scholars’ lack of interest in regional language variation did not, of course,
mean that they were not familiar with the phenomenon. On the contrary, ancient,
medieval, and early Italian Renaissance authors were sufficiently aware that a
language could show local particularities, and that two closely related speech
forms could be considered to constitute one language. This awareness did not
limit itself to Latin. Two well-known examples are the Roman historiographer
Tacitus (c.  55–c.  120) and the Early Christian author Jerome (c. 347–419).
Tacitus described the speech of the Bretons and Gauls as ‘by no means very
different’.¹² Jerome, on the other hand, noted that Galatian, spoken in Anatolia
(in present-day Turkey), and Trierisch, spoken in the German city of Trier, were
¹² De vita Iulii Agricolae 11.3: ‘sermo haud multum diversus’.

.    25
‘almost the same language’, albeit with some corruptions in the former.¹³ Tacitus
and Jerome were no doubt right in their observations on these tongues, which all
belong to the Celtic family of languages according to modern insights.
Gradually, the spectrum of languages in which regional variation was noticed,
broadened in the Latin West. In the later Middle Ages, the chronicler Gerald of
Wales (1146–1220) was able to remark differences in Welsh and English, and to
draw a comparison between the two linguistic contexts. Gerald was, however, still
struggling to express in words the two levels of variation he noticed: that of Welsh
and English, on the one hand, and that of Welsh and English regional varieties, on
the other. The latter he tried to label, rather unsuccessfully, as proprietas idiomatis,
a tautological phrase meaning something like ‘property of particularity’ (Gerald of
Wales 1868: 177). Much like the ancient Romans, Gerald was aware of different
contexts of regional language variation but proved himself unable to develop
clear-cut concepts to chart this phenomenon.
The difficulty of designing adequate conceptual tools to deal with regional
language variation was partly due to the limited reflection on linguistic diversity
in the period before 1500. Scholars were more interested in Latin, a well-
established language with a fixed form. It is telling that only Latin was taken
into account by modistic grammarians, an influential school of language philo-
sophers active in thirteenth-century Paris and aiming to uncover the very foun-
dations of grammar.¹⁴ The inability to come up with workable concepts was partly
also a consequence of the linguistic realities many medieval writers faced. In a
time when the vernacular languages were not yet being upgraded and standard-
ized as full-fledged written tongues, it was often difficult for scholars to perceive
clear borders between cognate and neighbouring languages and their varieties.
The Germanic tongues in Western Europe, for instance, were believed to consti-
tute one long chain of related tongues. The Italian humanist Enea Silvio
Piccolomini (1405–1464), later known as Pope Pius , offered a wonderful
example of this point of view. Shortly after 1431, Piccolomini described the
wide extension of Germanic tongues in terms of a concatenatio, a linguistic
sequence ranging from Austria to Scotland—a dialect chain, a present-day linguist
might be tempted to say.¹⁵ Indeed, pre-1500 scholars usually found it far from
obvious and often simply irrelevant to distinguish between different types of
linguistic diversity.
A well-known exception is Dante Alighieri (c. 1265–1321), who composed a
treatise On vernacular eloquence around 1305. In this Latin work, the Tuscan poet
¹³ Commentarii in  epistulas Paulinas, Ad Galatas 2.3: ‘lingua eadem paene’. See Denecker (2017:
238, 243–4).
¹⁴ On modistic grammar see e.g. Pinborg (1967); Bursill-Hall (1971); Dahan et al. (1995: 266).
¹⁵ Piccolomini (1960: 350): ‘Scotus vero, quantum ego meis fatis in eam plagam deductus perpendi,
non plus ab Anglico quam Australis a Bavaro distat, ut hanc Theutonicam linguam nexu et concate-
natione quadam ab Austria usque in Scociam facile productam videas’. See e.g. Borst (1957–63: 969).
26      
first distinguished the lingua sì, Italian, from the lingua oïl and the lingua oc, the
Romance tongues of northern and southern France, respectively. These tongues
were, Dante claimed, initially unitary but had been diversified as an idioma
trifarium, a ‘threefold idiom’, by his time (Tavoni 1987: 425–8; 1990; Consani
1991: 159; Trabant 2006: 67–8). He proceeded by dividing the sheer infinite
variation in Italian into two main classes, which, in turn, were each characterized
by several different linguistic layers. Under these two main classes, there were at
least fourteen ‘primary variations’ (variationes primae) across different Italian
regions. These fourteen varieties fell into numerous ‘secondary variations’ (var-
iationes secundariae), spoken in smaller regions and cities. Dante went even
further by introducing variations that were ‘subsecondary’ (subsecundarius) and
by arguing that there were a thousand different tongues within cities. What is
more, there were even differences per family (De vulgari eloquentia 1.10.4, 1.10.9,
and 1.19.3).
Apart from dissecting the phenomenon of linguistic diversity, Dante’s ideas are
of interest for another reason, too. They suggest that medieval scholars focused
principally on the opposition between Latin, on the one hand, and the chaos of
vernacular varieties, on the other, or in Dante’s terminology: grammatica versus
vulgaria. The Florentine poet’s interpretation of this opposition was, however,
rather atypical, since he valued the natural vulgaria more highly than artificial
grammatica. He even promoted an ‘illustrious’ form of the vernacular, a vulgare
illustre, as an alternative for Latin grammatica.
In Quattrocento Italian humanism, the opposition Latin/vernacular took on a
new dimension, as it gave rise to a debate over their precise relationship.¹⁶ There
were two main clashing opinions, to present the matter in a somewhat simplified
manner. A first group of scholars, including Biondo Flavio (1392–1463) and
Poggio Bracciolini (1380–1459), believed that the elite and the populace spoke
the same language, albeit with slight, socially determined differences, whereas
another group of authors, including Leonardo Bruni (c. 1370–1444) and Lorenzo
Valla (c. 1407–1457), claimed that the Roman lower classes employed a language
entirely different from the Latin of the higher classes. It can be readily seen how
this debate had implications for views on the relationship between Italo-Romance
varieties and Latin. The different degrees of linguistic diversity were barely
relevant in a model focusing on the opposition Latin/vernacular, if at all. It was
largely an either/or question: did these linguistic forms constitute either one
single language or entirely distinct languages? In other words, these scholars had
no use for an intermediate concept like dialect in the linguistic themes they
debated.
¹⁶ See e.g. Tavoni (1982, 1984); Mazzocco (1993: 13–105); Fausel et al. (2002: 200–7); Coseriu and
Meisterfeld (2003: 149–71); Celenza (2009); Eskhult (2018).
.  27
In sum, it is difficult to maintain on the basis of the available evidence what

medievalist Rüdiger Schnell (2008: 568–9) has contended for the later Middle
Ages on the basis of German examples, namely that dialects were supposed to vary
merely phonetically, whereas languages differed not only as regards their sounds
but also lexically, morphologically, and syntactically. Such a generally accepted
language/dialect distinction cannot be posited for late medieval times. One could
assert, at best, that a number of late medieval authors were, usually very vaguely,
aware of the fact that languages differ in regional terms and that linguistic
variation comes in different degrees.
2.5 Conclusion
Ancient, medieval, and early Renaissance references to, and conceptions of,
regional language variation usually had a very incidental character, revealing
different levels of awareness of this linguistic phenomenon. For the history of
the conceptual pair, Greek interpretations of the key term diálektos are of central
importance. At an early stage, this word was tied to notions of ethnicity, locality,
and, especially, literariness, since it was first and foremost associated with the
study of Greek texts and their diversified linguistic appearance. The philological
perspective of ancient and medieval Greek scholars dominated and was bound to
leave an indelible mark on later ideas, as I will argue in Part II. For this reason,
I have concentrated on the Greek tradition, with some attention to the way Greek
conceptions were received in the Latin West and other idiosyncrasies of Western
medieval thought.
It would be a fruitful avenue for further research to analyse how early
approaches to dialectal variation relate to each other cross-culturally, but for my
purposes it will suffice to conclude the following on the period before 1500: even
though there was widespread awareness of regional language variation, and some
authors such as Dante Alighieri even knew that linguistic diversity comes in
different degrees, most scholars did not make a clear-cut conceptual distinction
between a language-like and a dialect-like entity. There was, however, one con-
spicuous late medieval exception, which fully deserves a separate treatment.
3
The exception to the rule
Lingua and idioma in Roger Bacon’s thought
Roger Bacon (c. 1214/20–c. 1292) was a man of many firsts.¹ Bacon was, for
instance, most likely the first European to describe gunpowder, and a pioneer in
optics, a discipline he aspired to introduce into the quadrivium of the new
universities in Western Europe, such as the one in Paris, where he lectured for
quite some time. In this and other domains of study, Bacon was, moreover,
innovative in adopting an inductive approach rather than following the deductive
methods of his colleagues (cf. Figure 3.1). He attached great importance to
experimental and empirical research in order to reach more general conclusions.
It was this very desire to start from concrete reality that led him to become an early
advocate of an ad fontes approach in theology, which, in his view, should focus on
the Bible text. For this reason, he emphasized the importance of mastering
‘wisdom languages’ (linguae sapientales) other than Latin, too: especially Greek,
Hebrew, Arabic, and Aramaic. Indeed,
knowledge of languages is the first gate to wisdom, and especially among the
Latins, who do not have the text of theology or philosophy, unless from foreign
languages; and that is why every man should know languages.²
Or in a more poetical fashion:
[The idiomatic individuality of every language] was the main reason why all
saints, philosophers, and ancient sages wanted to know other languages, so that
they could drink the waters of wisdom more sweetly and more fully in the source
itself.³
Bacon even went a step further by strongly recommending the new universities
to include the wisdom languages in their curricula and to stop focusing exclusively
¹ Information on Bacon’s life and works is mainly based on Molland (2004) and Hackett (2013).
² Bacon (1859: 102): ‘Notitia linguarum est prima porta sapientiæ, et maxime apud Latinos, qui non
habent textum theologiæ, nec philosophiæ, nisi a linguis alienis; et ideo omnis homo deberet scire
linguas’. See also Bourgain (1989: 318–19).
³ Bacon (1859: 90): ‘fuit causa principalis quare sancti omnes, et philosophi, et sapientes antiqui,
voluerunt scire linguas alias, quatenus in ipso fonte dulcius et plenius biberent aquas sapientiæ’.
DOI: 10.1093/oso/9780198845713.001.0001
     29
Figure 3.1 Roger Bacon at his Merton College observatory, oil painting by Ernest
Board
Source: Wellcome Collection. CC BY 4.0
on Latin. It therefore comes as no surprise that he wanted to contribute to the

study of the wisdom languages by composing grammars of Greek and Hebrew, a
highly exceptional feat in the thirteenth-century West. Whereas his unique Greek
30     
grammar has been preserved in its entirety, only a fragment of his Hebrew
handbook seems to survive, and it is unclear whether he ever finished it (Hirsch
in Bacon 1902: 201).
3.1 Bacon at Babel
Let me take a closer look at Bacon’s linguistic horizon and his motives to linger at
length on linguistic diversity. His perspective was still very much Latin-centred,
since Latin remained the most familiar scholarly language for him and his
colleagues (Dahan et al. 1995: 266). Bacon pointed out, however, that the language
of the Romans was in various respects indebted to Greek and Hebrew, languages
which he no doubt sensed to be much older and which he described in contrast to
Latin in his pioneering grammars (Hovdhaugen 1990: 117). As can be expected
from a thirteenth-century intellectual, Latin was the only language in which he
composed his scholarly works. Yet unlike most of his colleagues, he did not blindly
accept the monopoly of the language of the ancient Romans on science. He was well
aware that there were other languages of interest for scholarship, most importantly
Greek and Hebrew. To this end, he had mastered both these languages, making him
an exceptional medieval instantiation of Jerome’s ideal of the vir trilinguis, the
‘trilingual man’. Bacon’s linguistic mastery did not stop there, however, as he also
had some notions of Biblical Aramaic and, to a lesser extent, Arabic.
Bacon’s interest in the three so-called sacred languages must have been part of
a broader tendency in Franciscan circles in late medieval times. The Flemish
Franciscan scholar Gerard of Huy, for instance, was the author of the work
Triglossos, id est liber trium linguarum, basically a glossary, in verses, of Latin,
Greek, and Hebrew words in the Bible (Roest 2015: 97). Another example is William
de la Mare, an English Franciscan friar active in the 1270s in Paris and influenced
by Bacon, who composed a glossary of Hebrew and Greek words in the Bible and
who was concerned with the correctness of the Vulgate text (Marenbon 2004).
The fact that the Ancient Greek dialects had acquired literary status and thus
had become part of the grammatical canon of Greek forced Bacon to comment on
this matter in his Greek grammar. He drew up this work, which followed in the
tradition of Byzantine grammar and is the first extant Greek grammar written in
Latin, around 1268. Still, his grammar was much more concerned with the Greek
dialects than the Byzantine elementary handbooks on which he could rely and
which were built up in question-and-answer format, and many of Bacon’s remarks
were idiosyncratic (see also Vandewalle 1929: 54–5; Consani 1991: 151; Rosier-
Catach 1997: 86).
Most of Bacon’s observations on dialectal diversity are found in his Greek
grammar, but he also treated the theme of linguistic diversity in his Major work
.     ? 31
of about 1266–7 (cf. Formigari 2004: 84). This work was intended as a lengthy
synthesis of natural philosophy, written in response to a mandate of Pope Clement
 (ruled 1265–8), of which Bacon learned in July 1266 while in Paris. Bacon also
treated this theme in his Third work, most likely conceived as an adapted shorter
version of his Major work, which ran the risk of getting lost during its transpor-
tation to the Holy See. Finally, in his Compendium of the study of philosophy
(c. 1271), a critical description of the intellectual situation in Paris with a defence
of his own work, he also elaborated on the issue of linguistic diversity and dialectal
variation.
Bacon did not restrict himself to noticing variation in Greek, as he showed
himself aware of diversity within vernacular tongues as well. Even though the
wisdom languages were the focus of his attention, Bacon had an ear for the
tongues of his contemporaries, too, with which he came into contact on his travels
to different parts of Europe. Born in Somerset, where a South Western variety of
Middle English was spoken, he taught at Oxford, where the South Western, West
Midland, and East Midland dialect areas came together, and at Paris, to which
students came from every area of France. Under these circumstances he must
have quickly found out that vernacular languages exhibited regional differences
much like those of Greek (cf. Hofmann 1883; Bourgain 1989: 325). All in all,
however, Bacon was not very enthusiastic about these vulgar tongues, which in
his eyes were useless for learning, since they lacked the necessary philosophical
terminology (Dahan et al. 1995: 267). Yet his interest in all things empirical
nevertheless seduced him to comment on this phenomenon, undeniably present
in everyday life.
3.2 Another first for Bacon?
In his attempts at mapping out regional linguistic diversity, Bacon introduced a

distinction that helped him make sense of this phenomenon. As he wrote very
quickly and very copiously, which resulted in frequent repetitions throughout his
writings, we are relatively well-informed about his ideas on this matter. Indeed, by
putting together the repeated statements, which differ in detail, we can build up a
richer picture than we would get from any one statement (cf. Bourgain 1989: 318;
Fausel et al. 2002: 196).
When discussing the varieties of the Greek language in his Greek grammar,
Bacon proposed to carefully distinguish between two types of linguistic entities,
which he called lingua and idioma:
For an idiom is a specific property of a language, which one people uses

according to its custom. And another people uses another idiom of the same
32     
language. For idion in Greek is proprium in Latin, whence idioma, i.e. proprietas.⁴
And I call these diversities idioms and not languages, as many do, because, in
reality, they are not different languages but different properties which are idioms
of the same language.⁵
An idiom is a property of speaking in a language.⁶
It immediately leaps to the attention that Bacon was operating with linguistic
concepts on two different levels. There were different ‘languages’ (linguae), and
each of these ‘languages’ was internally diversified in terms of ‘idioms’ (idiomata).
This strict division, Bacon suggested in passing, was an innovation of his own.
By means of the parenthetic remark ‘as many do’ in the second passage, he
complained about the carelessness of other scholars, who used both terms
interchangeably. Yet Bacon was not always straightforward himself. His numer-
ous definitions of idioma and his usage of the term reveal a confused concept
(cf. Rosier-Catach 1997: 85). An idioma was interpreted as a particularity of
speech within one and the same lingua, that is to say, with clear subsumption of
idioma under lingua. This relationship most obviously emerges from Bacon’s
repeated usage of the genitive phrase idiomata eiusdem linguae, ‘idioms of the
same language’.
Yet Bacon seems to have applied the term idioma not only to native varieties of
a language such as Burgundian French and Northern English, but apparently also
to different orthographies and pronunciations of Latin, for instance, by German
and Spanish natives. For he stated:
Indeed, in the Latin language, which is one, there are many idioms. For the
substance of the language itself consists in these things in which all clerics and
scholars take part. But there are many idioms according to the multitude of
nations using this language. For in many respects the Italians pronounce and
write it in one manner, the Spanish in another, the French in yet another, the
Germans in still another, and the English, too, in another, and so on. So in this
way there was among the Greeks one language according to substance, but there
were many properties.⁷
⁴ Bacon (1902: 26): ‘Idioma enim est proprietas lingue determinata, qua vna gens vtitur iuxta suam
consuetudinem. Et alia gens eiusdem lingue vtitur alio idiomate. Idion enim grece est proprium latine, a
quo idioma hoc est proprietas’.
⁵ Bacon (1902: 27): ‘Et voco has diuersitates idiomata et non linguas vt multi vtuntur, quia in veritate
non sunt lingue diuerse sed proprietates diuerse que sunt idiomata eiusdem lingue’.
⁶ Bacon (1902: 75): ‘idioma est proprietas fandi in aliqua lingua’.
⁷ Bacon (1902: 26–7): ‘In lingua enim latina que vna est, sunt multa idiomata. Substancia enim
ipsius lingue consistit in hijs in quibus communicant clerici et literati omnes. Idiomata vero sunt multa
secundum multitudinem nacionum vtencium hac lingua. Quia aliter in multis pronunciant et scribunt
ytalici, et aliter hyspani, et aliter gallici, et aliter teutonici, et aliter anglici et ceteri. Sic igitur fuit apud
grecos vna lingua secundum substanciam sed multe proprietates’.
.     ? 33
There is no doubt that Bacon was indeed referring to the diversified spelling and
pronunciation of the Latin language by speakers of different vernacular tongues
rather than stating that Italian, Spanish, French, German, and English were
daughter languages of Latin.⁸ The former view is corroborated by the fact that
Bacon still regarded Latin as the living mother tongue of Western Europe. After
all, this idea implies that Latin was subject to internal variation in a way similar to
other Western European vernacular tongues and to Greek, another ‘wisdom
language’. Also, Bacon was not only insisting on variation in Latin pronunciation
but also in writing. He was apparently thinking of the written language par
excellence, Latin, and not of the so-called vulgar languages, which lacked a
standard as well as an established written tradition.
Entirely in line with the Greek tradition, Bacon stressed the connection between
an idioma and a natio, ‘tribe’ or ‘nation’: ‘For an idiom is a property of a language
with a certain nation’.⁹ For the sake of convenience, the Latin term natio will be
translated in this book as ‘nation’, but it should be consistently interpreted in the
non-political sense of ‘group of people (allegedly) sharing the same ethnic back-
ground and living in the same area’, unless otherwise indicated. Bacon was,
moreover, aware of the fact that speech diversifies over space just like everything
else that is human, and for this dialect–space correlation he might have been
indebted to the Greek tradition, too (e.g. Bacon 1859: 438–9; see Lusignan 1986:
71–2; Bourgain 1989: 318).
Bacon did not explicitly define the term lingua but used it in various senses. To
the English scholar, it designated not only individual languages such as Latin, but
also groups of genealogically related varieties which were considered to constitute
one lingua. For instance, he regarded all Gallo-Romance tongues as one lingua
Gallicana (Fausel et al. 2002: 197). Bacon nonetheless offered a glimpse of his
conception of lingua when discussing the Greek language in his grammar. In his
opinion, the lingua Graeca consisted in those elements ‘in which all Greek nations
partook, and these elements are called “common” ’.¹⁰ The source for this view of
his must have been Byzantine scholarship, in which the idea that the Koine
consisted of what was common to the dialects was widespread. For instance,
John the Grammarian recalled that some scholars contended that the Koine had
no particularities of its own, but was compounded out of the four other dialects
(Manutius et al. 1496: 236).
Bacon was also convinced that related idioms showed only superficial, acci-
dental variation, for instance, in terms of pronunciation, but constituted a unity in
⁸ On this point I agree with Bourgain (1989: 326), Rizzo (1990: 37–40), and Rosier-Catach (1997: 85),
while disagreeing with Consani (1991: 148–56).
⁹ Bacon (1900: 73–4): ‘Idioma enim est proprietas linguae apud aliquam nationem determinatam’.
¹⁰ Bacon (1902: 27): ‘Natura igitur ipsius lingue grece consistit quantum ad ea in quibus omnes
naciones grece communicabant et hec vocantur communia’.
34     
substance.¹¹ This assertion must be viewed against the background of Aristotelian

ontology. It is inevitable to quote in this context Bacon’s notorious comment on
grammar, which was likewise informed by Aristotelian concepts, but remains
difficult to interpret: ‘grammar is one and the same in terms of substance in all
languages, even though it can vary accidentally’.¹² Bacon’s phrasing here reminds
of the later doctrine of speculative grammar that was only taking a start around his
time of writing, but his ideas must be set apart from them, as Even Hovdhaugen
(1990: 124–5) has convincingly argued by comparing the terminology of Johannes
Dacus’s Grammatical overview (Summa grammatica) of c. 1280 with Bacon’s.
Also according to Hovdhaugen (1990: 127), the sentence should be interpreted as
follows: ‘The science of analysing and describing language is [in essence] one and
the same for all languages although there are accidental variations’. This inter-
pretation seems indeed very plausible, since Bacon probably conceived of gram-
matica as the categories of grammar that were largely universal, with some
accidental particularities. What is more, he explicitly emphasized the utility of
comparing Greek and Latin grammar, which did not differ substantially in his
eyes. In fact, Latin was derived from Greek grammar, Bacon asserted on the
authority of Priscian, a favourite source of his. He did not refer to a specific
passage in Priscian’s Latin grammar, but he was surely thinking of comments of
this early sixth-century grammarian on the close connection between Latin and
the Aeolic Greek dialect. Indeed, Priscian, too, was convinced that Latin derived
from Greek and especially from Aeolic—‘Aeolism’ in short, which in antiquity was
more a commonplace than a true theory (Stevens 2006–7). Priscian principally
related this affinity to the Latin letter <F> and the formally identical Aeolic
digamma, a grapheme representing the sound [w] (Keil 1855–80: , 15, 1–6 and
, 17, 10–17). The assumption that grammatical categories were universal also
seems to emerge from the fact that Bacon (1902: 27, 207) linked the Greek and
Hebrew articles with the Latin ‘articular pronoun’ (pronomen articulare), by
which he meant demonstrative pronouns such as hic.
It is confusing that Bacon (1902: 26–7) used a highly similar expression to
describe regional linguistic variation in the vicinity of his notorious dictum on
grammar. In particular, he contended that Latin and Greek, although one in
substance, were each characterized by many ‘idioms’ (idiomata/proprietates).
The precise relationship of this statement to the passage just quoted remains
somewhat unclear, but it seems justifiable to argue that they should not be read
in close connection, even though they were both shaped by Aristotelian concepts
¹¹ Bacon (1859: 90): ‘Nam lingua Latina est in his omnibus una et eadem, secundum substantiam,
sed uariata secundum idiomata diuersa’.
¹² Bacon (1902: 27): ‘grammatica una et eadem est secundum substantiam in omnibus linguis, licet
accidentaliter uarietur’. See e.g. Bursill-Hall (1975: 201), Maierù (1983: 743–4), Lusignan (1986: 67),
Bourgain (1989: 320–1), Hovdhaugen (1990), Rosier-Catach (1997: 85), and Formigari (2004: 84) for
suggestions of interpretation.
.       35
and appear in close vicinity. In short, the phrase on grammar concerned the
quasi-universal nature of grammatical categories, whereas the assertions on Greek
and Latin were, in fact, applications of his lingua/idioma distinction rather than
generalizing statements on the nature of grammar.
Bacon’s choice for the term lingua is no surprise, but why did he opt for the
Greek loanword idioma, from idíōma (ἰδίωμα), to denote a variety of a language
and not for dialectus, the Latinized form of a central term in Greek grammatical
tradition? As a matter of fact, he did use the Greek term diálektos as well as
dialectos, its transcription into the Latin alphabet, but only very occasionally and
exclusively to talk about varieties of Greek. Latinized dialectus, however, had not
yet been incorporated into the metalinguistic terminology of Bacon’s days (see
Chapter 2, Section 2.4). He thus principally relied on other terminological options
to express concepts related to linguistic variation. Apart from idioma, he also
referred to dialects and dialectal variation in terms of proprietas (loquendi),
diversitas, and variare. These words were all well-established Latin expressions.
Idioma, for instance, had been in constant use since late antiquity as a word
denoting, among other things, linguistic entities, especially in contexts in which
their particular properties were stressed.
3.3 An outlook on different languages and their dialects:

a central precondition
How and why could Bacon develop a conceptual scheme of individual languages
each subsuming several different idioms? In his Major work, he explained the
opposition lingua/idioma by means of the example of France, where he wrote
this work:
but it is impossible that the property of one language is preserved in another. For
also idioms of the same language are varied among diverse [nations], as is clear
from the French language, which is varied by a manifold idiom among the
Walloons, the Picards, the Normans, and the Burgundians. And what is properly
said in the idiom of the Picards, becomes rough among the Burgundians, indeed
even among the Walloons, who live nearer to Picardy; so how much more will
this occur among different languages? Therefore, what is well done in one
language, cannot be transferred to another according to the property it had
from the start.¹³
¹³ Bacon (1900: 66–7): ‘sed impossibile est quod proprietas unius linguae servetur in alia. Nam et
idiomata ejusdem linguae variantur apud diversos, sicut patet de lingua Gallicana, quae apud Gallicos
et Picardos et Normannos et Burgundos multiplici variatur idiomate. Et quod proprie dicitur in
idiomate Picardorum horrescit apud Burgundos, immo apud Gallicos viciniores: quanto igitur magis
accidet hoc apud linguas diversas? Quapropter, quod bene factum est in una lingua, non est possibile ut
36     
Here, Bacon was explaining the idiomatic characteristics of different linguistic

forms in the context of translation theory. All idioms of a language had their
idiosyncrasies, which applied a fortiori at the level of languages and made literal,
word-for-word interpretation impossible. In this passage, we can find a hint as to
why Bacon adopted idioma as his term to denote varieties of one and the same
language. Even though they showed many similarities, it was their mutual differ-
ences and idiosyncratic ‘properties’ (idiomata) that immediately caught the eye.
Bacon was, in other words, seemingly jumbling together two concepts modern
linguists tend to clearly distinguish: dialect and idiom.
The idiomatic properties of each variety of a language may, Bacon obser-
vantly remarked, cause mutual laughter. In his Third work, he rephrased his
views on this issue and expanded his perspective so as to include almost the
entire Romance area:
for an idiom is a property of a language distinct from another, such as the Picard,
the Walloon, and the Provençal, and all idioms from the borders of Apulia to the
borders of Spain. For the Latin language is in all these [regions] one and the same
in terms of its substance, but varied according to different idioms. And we
manifestly see that something sounds excellent and proper in one idiom and
ridiculous in another, as is clear not only from those far removed but also from
those very close-by, e.g. from the Picards and the Walloons, for they laugh at
each other.¹⁴
It can be noted that Bacon was here clearly thinking of vernacular Romance
variation rather than of variation in the pronunciation and orthography of the
Latin language.
Bacon most likely abstracted the lingua/idioma distinction from the linguistic
contexts with which he was most familiar, most notably English and French, and
used it to describe and analyse these same as well as other linguistic contexts less
familiar to him, most importantly Greek and its dialects as well as the family of
languages today known as Semitic. This cognitive process is excellently exempli-
fied by his observation on the relationship between Hebrew and Aramaic:
transferatur in aliam secundum ejus proprietatem quam habuerit in priori’. Gallicus seems to refer to
the Walloons, whereas Gallicanus is more generic, denoting Gallo-Romance as a whole. For Bacon’s
ideas on French dialectal variation see also Hofmann (1883), Meyer-Lübke (1908), and Lusignan
(1986: 67–72).
¹⁴ Bacon (1859: 90): ‘nam idioma est proprietas alicujus linguæ distincta ab alia; ut Picardicum, et
Gallicum, et Provinciale, et omnia idiomata a finibus Apuliæ usque ad fines Hispaniæ. Nam lingua
Latina est in his omnibus una et eadem, secundum substantiam, sed variata secundum idiomata
diversa. Et manifeste videmus quod aliquid optime et proprie sonat in uno idiomate et ridiculose
sonat in alio; ut patet non solum de longinquis sed propinquissimis; sicut de Picardis et Gallicis; nam
mutuo se derident’.
.       37
And it is certain that the Hebrews and Chaldeans have the same language but a
different idiom, just like the Walloon and the Picard . . . the Hebrew says Eloim
for “God” or “Gods”; the Chaldean says Eloa for “sky” or “skies”. For “not”, the
Hebrew says lo, the Chaldean says la, and similarly in other words.¹⁵
In his Compendium of the study of philosophy, Bacon made a similar point, but
now in more general terms, stating:
for Aramaic and Hebrew speech differ as idioms of one language, just like the
Picard and the Norman, the Burgundian, the Parisian, and the Walloon; for there
is one language among all, namely the French language, but still in several parts it
varies accidentally; a diversity that makes idioms, not different languages.¹⁶
Other language families were described in similar terms. In the fourth part of
his Major work, devoted to geography, Bacon briefly commented on the Slavic
family of tongues, which he apparently regarded as idioms of one and the same
language, even though he did not use the term in this context:
The Russians are Christians and at the same time schismatic, as they have the rite
of the Greeks. Yet they do not use the Greek language, but rather the Slavonic
language, which is one of the languages that cover multiple regions. For it
comprises Russia, Poland, and Bohemia, and many other nations.¹⁷
It is conceivable that Bacon thought Russian, Polish, and Bohemian to be idioms

of one Slavonic language, as he linked them to different Slavic nations, even
though Pascale Bourgain (1989: 318) is right in emphasizing that Bacon was not
familiar with the Slavic tongues and suggests that he simply believed Slavic to be a
unitary tongue. Still, Bacon thought of regional variation as being inherent in
language, and even though he did not know any Slavic tongue, it went without
saying for him that a language spoken in multiple regions and by multiple nations
underwent diversification.
Bacon’s acquaintance with several dialect contexts was without doubt an
important trigger for his designing a lingua/idioma distinction. As a matter of
¹⁵ Bacon (1900: 73–4): ‘Et certum est quod Hebraei et Chaldaei eandem habent linguam, sed
diversum idioma, sicut Gallicus et Picardus . . . Hebraeus dicit Eloim pro Deo vel Diis; Chaldaeus
dicit Eloa, pro coelo vel coelis. Pro non, Hebraeus dicit lo, Chaldaeus dicit la, et sic in aliis’.
Cf. Demonet (1992: 27); Dahan et al. (1995: 275).
¹⁶ Bacon (1859: 438–9): ‘Chaldæus enim sermo et Hebræus differunt sicut idiomata unius linguæ; ut
Picardicum et Normanicum, Burgundicum, Parisiense, et Gallicum; una enim lingua est omnium,
scilicet Gallicana, sed tamen in diversis partibus diversificatur accidentaliter; quæ diversitas facit
idiomata non linguas diversas’.
¹⁷ Bacon (1900: 360): ‘Rusceni sunt Christiani et sunt schismatici, habentes ritum Graecorum, sed non
utuntur lingua Graeca, immo lingua Sclavonica, quae est una de linguis quae plures occupant regiones.
Nam Rusciam, Poloniam, et Bohemiam, et multas alias nationes tenet’. Cf. Rosier-Catach (1997: 84).
38     
fact, such a conceptual pair made most sense to a scholar engaged, first, in
separating different languages from each other and, second, in mapping out
their respective internal diversity. As I have contended at length, such an interest
clearly emerges from Bacon’s works. I can sum up here once more the different
linguistic contexts he mentioned, truly impressive for a thirteenth-century
scholar: English, French, Greek, Latin, Romance, and Semitic. Tellingly, Bacon
claimed that ‘whoever is happy with their property of speaking, while ignoring the
properties of speech of others, is a person of narrow learning (idiota)’.¹⁸ Bacon was
adapting here a passage in the Venerable Bede’s commentary on Acts and
projecting it onto the level of idioms of a language. Bede’s original reads as follows:
For persons of narrow learning (idiotae) were called those who, satisfied with
only their own language and with natural science, did not know the study of
letters, since the Greeks say “proper” as “ídion”.¹⁹
An outlook on multiple languages and their respective internal diversity, then,

seems to have been a necessary precondition for distinguishing between language
and dialect-like concepts. That it was not a compelling one has been established by
means of the example of the ancient Romans in the previous chapter. Lacking the
comparative-contrastive perspective of Roger Bacon, they were blindsided by their
assumption that the Greek literary dialects were something completely different
from regional variation within Latin.
3.4 Thomas Aquinas: a like-minded exegete?
Bacon was in good company in distinguishing two levels of linguistic entities that
exhibited different degrees of variation. There is a close parallel to his opposition
of lingua to idioma in the exegetical work of the Aristotelian philosopher and
Dominican friar Thomas Aquinas (c. 1225–74). In a commentary on the Book of
Psalms, Thomas was very careful in distinguishing loquela from sermo, stating
that ‘loquelae signify principal languages, but sermones signify varieties of idioms
in the same language’ or—put differently—‘loquelae are languages’, whereas
‘sermones are manners of speaking’.²⁰ His distinction between loquela and sermo
was triggered by an anacoluthic passage in Psalms 18.4 of the Vulgate version,
¹⁸ Bacon (1902: 26): ‘Et idiota qui est contentus sua proprietate loquendi, nesciens proprietates
sermonis aliorum’. Cf. Bacon (1902: 75).
¹⁹ Expositio actuum apostolorum 4.40: ‘Idiotae enim dicebantur qui propria tantum lingua natur-
alique scientia contenti litterarum studia nesciebant, siquidem Graeci proprium ἴδιον vocant’.
²⁰ In Psalmos reportatio 18.2: ‘Loquelae significant linguas principales; sed sermones significant
varietates idiomatum in eadem lingua. Vel loquelae linguae, sermones sunt modi loquendi’. See
Lusignan (1986: 61); Eloy (1998: 4).
.  :  - ? 39
which reads: ‘non sunt loquellae neque sermones quorum non audiantur voces
eorum’. ‘There are no languages, nor are there speeches of which the voices cannot
be heard’. This verse corresponds to Psalms 19.3 in the English Standard Version,
which translates the original Hebrew as follows: ‘There is no speech, nor are there
words, whose voice is not heard’.
Thomas’s terminological contrast resembles to some extent later oppositions of
dialect to language, even though he refrained from further elaborating on his
distinction, and his definitions, especially of sermo, remained quite vague. He did
seem to put the opposition into practice elsewhere in his oeuvre, when comment-
ing on the recognition of Peter as a Galilean by means of his speech in Matthew
26.73: ‘After a little while the bystanders came up and said to Peter, “Certainly
you, too, are one of them, for your accent betrays you” ’. In the Latin Vulgate
version, the last part reads: ‘loquella tua manifestum te facit’. Thomas Aquinas
attempted to explain the passage by rephrasing Jerome’s comments on it and by
referring to Gallo-Romance diversification, which was also Roger Bacon’s pre-
ferred example: ‘in the same language (lingua), there is often diverse speech
(locutio), as is clear in the Île-de-France, Picardy, and Burgundy, and still the
language (loquela) is one’.²¹ Again, Thomas used the term loquela here to refer to
language, just as in his commentary on Psalms, but its presence here is rather
curious. For the Vulgate passage of Matthew 26.73 has the Latin word loquela
refer to the Galilean characteristics of his speech, that is to say, to regional
variation within a language rather than to differences between distinct languages.
Thomas thus must have had the Psalms verse, and his interpretation of it, in mind
when commenting on Matthew’s gospel. For Thomas, lingua could apparently
serve as a synonym for loquela to refer to language, whereas he refrained from
using the term sermo here and opted instead for locutio to denote regional
diversity.
The example of Thomas indicates the importance of engagement with the Bible
text for eliciting observations on dialectal variation. The very same passage in the
gospel of Matthew had led Jerome, for instance, to argue that it pointed to
variation within Hebrew, as ‘every province and region had its properties and
could not avoid the native sound of speaking’—an explanation adopted by the
Venerable Bede.²² Jerome even linked Matthew 26.73 to the passage in Judges
12.5–6, which relates how two tribes of Israel, the Gileadites and the Ephraimites,
were waging war and the former succeeded in unmasking the latter by means of
linguistic evidence. The fleeing Ephraimites were unable to pronounce the word
shibbóleth, meaning either ‘stream’ or ‘part of a plant containing the grain’, and
²¹ Super Evangelium Matthaei reportatio 26.7.2296.1: ‘in eadem lingua saepe diversa locutio fit, sicut
patet in Francia, et Picardia, et Burgundia, et tamen una loquela est’.
²² Jerome, Commentarii in evangelium Matthaei 1.1452–6: ‘unaquaeque provincia et regio habebat
proprietates suas et vernaculum loquendi sonum vitare non possit’. Cf. Bede, In Marci evangelium
expositio 4.14.
40     
said sibbóleth instead, which led to their bloody execution.²³ The episode in Judges
was intensively commented upon, and shibbóleth would eventually lend its name
to the concept of shibboleth, designating a particular feature, often of a linguistic
nature, setting a specific group of people apart from others.
Some decades after Thomas Aquinas, another influential exegete, the
Franciscan Nicholas of Lyra (1265–1349), was reminded of Gallo-Romance vari-
ation in a manner similar to the great Aristotelian philosopher. Nicholas con-
trasted Picard to Parisian speech and emphasized that one’s speech variety could
be used as a distinctive mark revealing one’s geographical provenance, as a
shibboleth:
For even though the Hebrew language is one, it was still in some way varied
according to the diversity of provinces. And for this reason those from Galilee
spoke in some way differently from those who originated from Jerusalem, just as
the French tongue is one and yet those who are from Picardy speak differently
from those who live in Paris. And by such a difference it can be perceived whence
someone originates.²⁴
3.5 The Tower of Babel
One of the major linguistic events described in the Bible is the confusion of
tongues at the Tower of Babel in Genesis 11.1–9 (cf. Figure 3.2). The text, however,
does not offer much information on the precise nature of this confusion. The
reader is only told that mutual understanding was made impossible by God, so as
to bring the building of the Tower to an end:
Now the whole earth had one language and the same words. And as people
migrated from the east, they found a plain in the land of Shinar and settled there.
And they said to one another, “Come, let us make bricks, and burn them
thoroughly.” And they had brick for stone, and bitumen for mortar. Then they
said, “Come, let us build ourselves a city and a tower with its top in the heavens,
and let us make a name for ourselves, lest we be dispersed over the face of the
whole earth.” And the L came down to see the city and the tower, which the
children of man had built. And the L said, “Behold, they are one people, and
²³ See e.g. Swiggers (1981) for the Hebrew word, ‫ ִשׂ ֹבֶּלת‬/‫ ִשׁ ֹבֶּלת‬as opposed to ‫ִס ֹבֶּלת‬, in this passage. Cf.
Trabant (2006: 24).
²⁴ Nicholas of Lyra (1471: [81–2]): ‘licet enim linguua [sic] Hebrea sit una: tamen aliquo mo[do]
uariebat[ur] [sic] secundum diuersitate[m] prouinciarum. & ideo illi de Galylea aliter loquebant[ur]
aliquo modo: q[uam] illi qui erant de Hierusalem: sicut lingua Gallica est una: & tamen aliter loquunt
[ur] illi qui sunt de Picardia: q[ua]m illi qui habitant Parisius [sic]: & per talem uarietatem potest
percipi unde aliquis sit oriundus’. I quote from the printed editio princeps for lack of a modern critical
edition. See also Várvaro (2008: 224).
.     41
Figure 3.2 The Tower of Babel through the eyes of Pieter Bruegel the Elder, 1563
Held at the Museum Boijmans Van Beuningen, Rotterdam
Source: Wikimedia Commons; public domain
they have all one language, and this is only the beginning of what they will do.
And nothing that they propose to do will now be impossible for them. Come, let
us go down and there confuse their language, so that they may not understand
one another’s speech.” So the L dispersed them from there over the face of all
the earth, and they left off building the city. Therefore its name was called Babel,
because there the L confused the language of all the earth. And from there
the L dispersed them over the face of all the earth.
Some modern linguists might conclude that the absence of mutual intelligibility
among the workers indicates differences on the level of language rather than
dialect, but since the conceptual pair was yet to establish itself, this idea was
completely absent from ancient and medieval linguistic thought. Exegetes under-
stood that new languages were spoken, but most of them did not yet have the
conceptual apparatus to ask whether it was better to think of them as languages or
as dialects.
Still, one late medieval author, the Spanish scholar Rodrigo Jiménez de Rada
(c. 1170–1247), tried to understand the events at Babel by seemingly working with
42     
a distinction between languages and idioms in the first book of his History of
the vicissitudes of Spain, or Gothic history. As a consequence of the building of
the Tower, mankind was divided linguistically and, hence, also ethnically. In the
aftermath of this double separation, the tribes ‘distinguished the tongues by
idioms’.²⁵ Rodrigo did not explain what he exactly meant by these ‘idioms’ and
how they related to the various languages that emerged. He seems to have been
thinking of an intensification of the differences among the tongues that arose at
Babel rather than of an internal diversification within each of the tongues separ-
ately. That is, at least, how the use of the active verb may be interpreted, with each
tribe, driven by mutual hatred, looking to enhance mutual linguistic differences.
Two chapters further on, Rodrigo again juxtaposed lingua and ydioma, this time
with reference to Germanic-speaking areas. Germany, Denmark, Norway,
Sweden, Flanders, and England were said to ‘have a single language, even though
they are distinguished by idioms’.²⁶ The term ydiomata probably needs to be taken
here in the sense of ‘particularities’ or ‘properties’ rather than ‘separate linguistic
entities subsumed under one and the same language’ (~ dialects). In sum, Rodrigo
does not seem to have operated with a distinction between language- and dialect-
like entities, even though a superficial reading of his work may convey such an
impression.
3.6 Conclusion
I can only agree with Gilbert Dahan, Irène Rosier-Catach, and Luisa Valente when
they state that Roger Bacon was ‘one of the most subtle and original minds of the
Middle Ages, especially regarding the reflection on languages’.²⁷ For this reason,
Bacon is often bracketed together with Dante Alighieri in current scholarship.
Both late medieval thinkers were intrigued by the phenomenon of regional
variation within a language and formulated pioneering ideas about the matter,
albeit in clearly distinct contexts.²⁸ Bacon framed it within a broader insistence on
the variability of human culture, which he tied first and foremost to the factor of
space (Lusignan 1986: 71; Bourgain 1989: 318). In addition, he related regional
variation in a language to translation theory. Literal translation conveying all
nuances of the source language was impossible, since there existed idiomatic
differences even among varieties of individual languages.
²⁵ Historia de rebus Hispanie sive Historia Gothica 1.1: ‘linguas ydiomatibus distinxerunt’.
²⁶ Historia de rebus Hispanie sive Historia Gothica 1.3: ‘unicam habent linguam, licet ydiomatibus
dinoscantur’. Cf. Bonfante (1953–4: 681).
²⁷ Dahan et al. (1995: 267): ‘un des esprits les plus subtils et originaux du moyen âge, surtout pour ce
qui concerne la réflexion sur les langues’.
²⁸ See Hüllen (2001: 217–18); Fausel et al. (2002: 195). See Chapter 2, Section 2.4, for Dante’s views.
.  43
Despite the pioneering character of Bacon’s insights, most modern scholars

have only touched sideways on his lingua/idioma distinction. In response to
Even Hovdhaugen’s (1990: 117) call, I have endeavoured to cast a clearer light
on this issue, which is not as straightforward as is often maintained (cf. e.g.
Lusignan 1986: 68). By subsuming idioma under lingua, Roger Bacon was the
first to develop and define at some length a language/dialect-like distinction, an
innovation of which he was perfectly aware. He mainly related it to Aristotelian
notions, with related idioms only showing accidental differences, whereas lan-
guages differed substantially. Like diálektos in the Greek tradition, idioma was
also closely tied up with the entity of ‘nation’ (natio) in Bacon’s thought. It is
moreover noteworthy that the English scholar adopted an objective approach to
regional variation within English and French. He did not express any value
judgements on individual idioms, even though he was aware that idioms of one
language could give cause to mutual ridiculousness among speakers. Bacon’s
neutrality in this matter partly resulted from the linguistic context of thirteenth-
century Western Europe. Large-scale vernacular standardization was, after all,
still in the future, as were the prescriptive and subjective language attitudes that
went with it.
At the same time, Bacon’s interpretation of idioma remained somewhat con-
fusing, as he did not link it only to regional variation in vernacular tongues but
also to idiomatic properties causing untranslatability and to different ways of
writing and pronouncing Latin. Yet since he still conceived of Latin as a mother
tongue, this last idea probably seems inconsistent only from our present-day point
of view. Still, his interpretation of idioma was far from straightforward, and a
similar conclusion can be drawn for his vague usage of lingua, a term he did not
explicitly define.
Bacon’s case strongly suggests that the design of a language/dialect-like
distinction had a peculiar precondition, which could act as a triggering circumstance:
a familiarity with regional variation in two or more different languages. To scholars
with this specific kind of linguistic outlook, such a conceptual opposition made
obvious sense and was a useful descriptive tool for ordering and mapping out the
great linguistic diversity they noticed. The fact that Bacon was one of the few
medieval scholars who knew Ancient Greek well is surely no coincidence in this
regard. His linguistic horizon indeed was exceptionally wide, and most medieval
scholars lacked a panoramic view on language, a state of affairs that profoundly
changed only in the early modern era.
Even though linguistic variation was a well-known phenomenon during the
later medieval period, no generalizations on regional variation within a language
seem to be extant in texts written before about 1250. Even after this such
statements were rare, with the works of Thomas Aquinas and especially Roger
Bacon as the main exceptions. Bacon’s oeuvre remained, however, buried in the
back rooms of libraries and was only rediscovered and published in modern times.
44     
His Greek grammar, for instance, survives in four manuscripts but was published
only in 1902 (Nolan and Hirsch in Bacon 1902: lxv–lxxi; Consani 1991: 156;
Dahan et al. 1995: 274). Bacon thus remained an isolated early pioneer, whose
lingua/idioma opposition did not influence later linguistic thought. Instead, early
modern scholars independently reinvented a language/dialect opposition.
II
THE O RIGIN OF T HE
C O N C E P T U A L P A I R , 15 0 0 – 50
4
From dogs and hounds to
languages and dialects
The conceptual pair in Conrad Gessner’s work
Thus far I have treated of dogs those aspects that seemed common to
all dogs; now it remains for me to draw up those matters that pertain
to diverse dogs separately.¹
The sixteenth-century Swiss polymath Conrad Gessner (1516–65) was greatly

invested in cataloguing diverse entities, especially natural ones like animals and
plants. In his impressive publications, Gessner usually discussed the different
kinds in alphabetical order, even though his entries ‘often included related species
under the same heading’ (Ogilvie 2006: 216). So, for instance, he described dogs
first in general in his History of animals (cf. Figure 4.1). Then he pointed out that
they could be divided into different types, including hounds, the kind hunters
used in chasing prey. These types could in their turn be subdivided into different
subspecies. The range of natural entities the productive Zürich humanist tried to
catalogue and classify was not limited to animals and plants, however. He also
tried to impose order on the chaos of existing language forms in his much-read
Mithridates, the first language catalogue ever to be published. It was, however, in
his preface to Josua Maaler’s (1529–99) German–Latin dictionary of 1561 that he
voiced his language-classifying concerns most powerfully: ‘I do not separate Swiss
from German but make it a dialect thereof ’.² Indeed, just as a hound was a kind of
dog, Swiss was a kind of German; it was a dialect of it. Gessner’s active role in
relating these two forms of speech finds a striking expression in his phrasing. It is
his deliberate decision—facio, ‘I make’—to classify Swiss as a dialect of German.
Gessner’s observation presupposed a contrast between dialectus and lingua,
with the former subsumed under the latter, and suggests that, by his time of
writing, a distinction similar to Roger Bacon’s lingua/idioma opposition had been
introduced, a development of central significance thus far largely neglected by
modern scholars; a notable exception is Alinei (1980; cf. also Alinei 1984 [1981]:
¹ Gessner (1551: 235): ‘H de canibus dixi ea quæ canibus communia omnibus uidebantur:
reliquu[m] est ut quæ singulatim ad canes diuersos pertinent conscribam’.
² Gessner (1561: *.5): ‘ego Helueticam à Germanica non separo, sed eius dialectum facio’.
DOI: 10.1093/oso/9780198845713.001.0001
48        
Figure 4.1 Dogs in Conrad Gessner’s work

.     49
179). How and why did this evolution occur, and what evidence is available? And
in what ways did it differ from Bacon’s innovation? Before I attempt to answer
these questions both from a close-reading perspective in this chapter and by
means of a bird’s-eye view in the next, I should say a brief word on the early
modern language horizon and the linguistic profile of humanist scholars, indis-
pensable for an adequate understanding of the early sixteenth-century emergence
of the language/dialect pair.
4.1 Exploring the linguistic world
How narrow or expanded was the early modern linguistic horizon?³ There can be
no doubt whatsoever that it was much broader than it had been in ancient and
medieval times. The renowned intellectual historian Peter Burke (2004: 15–42)
has rightly spoken of ‘the discovery of language in early modern Europe’, even
though it might be more accurate to refer to languages in the plural. The
phenomenon of language had been on the scholarly agenda for centuries, but
early modern scholars were the first to explore at considerable length the world’s
linguistic plurality. This discovery of languages can be framed within the ‘early
modern information explosion’ and the concomitant ‘info-lust’ of scholars (Blair
2010: 11–12). They were overwhelmed by all kinds of new information in different
branches of learning, a development triggered first and foremost by the printing
press and European explorations of the world.
With which languages were sixteenth-century scholars primarily involved? To
begin with, for the first time in history, it was no longer a rarity to embody
Jerome’s ideal of the vir trilinguis, the scholar versed in the three so-called sacred
languages: Hebrew, Greek, and Latin, the languages in which the title ‘King of the
Jews’ was believed to have been inscribed on Jesus Christ’s cross. In the sixteenth
century, this ideal became institutionally embodied in the foundation of trilingual
colleges, including the Leuven Collegium Trilingue (1517) and the Paris Collège
Royal (1530; see e.g. Papy 2018). In the wake of Hebrew, other so-called Oriental
tongues excited increasing attention. Apart from Arabic and Syriac, two other
Semitic tongues, the Persian language, of Indo-European descent, also attracted
attention; it was even claimed to be closely akin to the Germanic branch of
tongues (Van Hal 2018). The explorations of Columbus and others moreover
brought new languages within the scope of European scholars, who were often
baffled by their entirely different appearance (cf. Figure 4.2). Their interest was
aroused mainly from a missionary point of view, with conversion of pagans as the
ultimate goal. To this end, many clerics did their best to draw up grammatical
³ For an outstanding overview of the early modern linguistic horizon see Percival (1992).
50        
Figure 4.2 World map by Abraham Ortelius (1527–98) showing the expansion of the
European horizon, 1570
Source: The Library of Congress. Public domain
codifications of languages such as Zapotec, a tongue spoken in what is now

Mexico, an endeavour they undertook with varying degrees of success.⁴
At home, the vernacular tongues, up to then little more than conglomerates of
related varieties, were in the process of being standardized. Intellectuals tried to
promote one dialect as the norm or to design a common, supraregional variety,
and to codify and elaborate it in grammars and handbooks. They aimed to arrive
at an efficient and elegant speech form that was widely used and adequate for
administrative as well as literary purposes.⁵ This new language could be linked to a
centralized state, as in France, where the upgraded vernacular was extensively
praised by Joachim du Bellay (c. 1522–60), but that was not necessarily the case, as
in Italy, where du Bellay’s source of inspiration, Sperone Speroni (1500–88), made
an urgent plea in favour of the vernaculars.
Scholars thus put into motion the complex process of standardization, which
occurred at different paces for each language. There were, for instance, strong
centrifugal tendencies in politically diversified areas such as Germany (Burke
2004: 109). Sometimes there was an active and extensive debate about the variety
that should serve as the basis for the common language (see e.g. Giard 1984: 46–7;
⁴ On the history of missionary linguistics see e.g. Hovdhaugen (1996); Zwartjes (2012);
Zimmermann and Kellermeier-Rehbein (2015).
⁵ For the four processes of standardization, selection, codification, elaboration, and acceptance,
Haugen (1966) is the classical reference. See e.g. also Joseph (1987).
.     51
1992: 206–9). The prototype of such a debate is the Italian language question
(questione della lingua), which has been the subject of intensive study (e.g. Tavoni
1984). In other cases, the selection of a specific variety occurred more or less
naturally as a result of the prestige of its speakers, as in the cases of South Eastern
English and Île-de-France French, or of other historical circumstances such as the
Reconquista in the case of Castilian Spanish (Burke 2004: 97). The elaboration of
one variety as the norm could, however, entail the admixture of elements from
other dialects. Standardization processes were catalysed by the commercialization
of the printing press, which almost naturally raised questions about the lack of
uniformity of vernacular tongues. The English printer William Caxton’s
(c. 1415–92) sighs about the variability of English, for instance, are well-known
(Caxton 1490: .i; Harris and Taylor 1997: 87–93; Van Rooy 2018b: 198). Apart
from issues of standardization, increased travel activities also fostered curiosity
about the regional variation of different languages.
Finally, early modern scholars expressed a keen interest in the extinct ancient
languages of Europe, such as Gaulish, the ancient Celtic tongue of what is now
France (see Van Hal 2013/14), and the then enigmatic tongue of the Gubbio
tablets, discovered in 1444 and identified by modern scholars as Umbrian, an
Italic tongue related to Latin. Their rediscovery forced intellectuals to reflect
extensively on issues of language classification. Especially Gaulish played a key
role in the constitution of the dialect concept in the sixteenth century, as I will
argue in the next chapter (Section 5.5).
This brief sketch has highlighted the multilingual perspective of early modern
Europe, and especially of its Republic of Letters (cf. Bots and Waquet 1997), in
which being a polyglot was considered an asset. I can even go a step further and
assert that a considerable number of scholars had what could be dubbed a
multidialectal outlook. For the first time in history, it was not uncommon for
an intellectual to be familiar with regional variation in two or more different
languages. It seems, by the way, better to call this perspective a ‘multidialectal
outlook’ rather than a ‘polydialect competence’ (by analogy to ‘polyglot’), since
it was highly unusual for a scholar to truly master different dialects of different
languages. The development of this new perspective stood in direct connection
with the Renaissance rediscovery of the Ancient Greek language and literature
and the subsequent anchoring of Greek studies in Western European scholar-
ship. As a matter of fact, the Greek language provided an additional dialect
context next to a scholar’s native one, triggering rudimentary and intuitive
comparisons (Van Rooy 2020a: Chapter 8). What is more, the fact that the
Greek dialects were the linguistic vehicles transmitting highly valued literary
texts left open the possibility that vernacular dialects could be a study object of
interest, too.
When exactly did the Greek dialects become known to humanist scholars? In
the first phase of the Renaissance interest in Greek, the prima grecità, the human-
ist pioneers Petrarch (1304–74) and Giovanni Boccaccio (1313–75) had tried to
52        
master Greek without much success. From the very end of the fourteenth century
onwards, Greek studies succeeded in gaining a firm foothold in the north of the
Italian peninsula, first in Florence and later also in other city-states. Numerous
Byzantine scholars migrated there to teach the elements of Greek to Western
students, an influx intensified by the Fall of Constantinople in May 1453. This
seconda grecità ended around the turn of the sixteenth century, when Greek
studies ripened and Western European scholars gradually took the wheel from
their Byzantine masters. The takeover resulted in the terza grecità, with Greek
studies spreading and booming in different parts of Western Europe, a develop-
ment greatly stimulated by the medium of the printing press, as Elizabeth
L. Eisenstein (1966: 46–7) has rightly suggested.⁶ Ever since, Western intellectuals
have evinced an abiding interest in the Greek heritage. What is more important to
my argument, however, is that only in this third phase did the Greek dialects
become an inherent part of the curriculum of the student of Greek literature.
In conclusion, the largely monolingual, Latin-focused culture of medieval
Western Europe gave way to the multilingual and multidialectal perspective of
the Renaissance, with humanist scholars enthralled by many different types
of tongues, which they categorized roughly as ‘ancient’, ‘Oriental’, ‘exotic’, and
‘vernacular’.
4.2 Conrad Gessner, certified cataloguer and compiler
The first scholar to make a serious attempt at systematically exploring the ever-
widening linguistic horizon was the Zürich humanist, physician, and compiler
Conrad Gessner (see Figure 4.3).⁷ Gessner did so by listing alphabetically and
describing, in his widely read Mithridates of 1555, a large number of languages
and dialects of the then known world, ancient as well as contemporary, while
sketching their mutual kinship along the way.⁸ His catalogue, written in Latin and
named after the ancient Pontic king Mithridates  (135–63 ), who allegedly
mastered more than twenty tongues (Figure 4.4), excellently embodies several
aspects of his scholarly activities. Gessner was first and foremost an eager com-
piler, for which he was partly motivated by financial considerations (cf. Springer
and Kinzelbach 2008: v, 39). But it was certainly not all money that drove Gessner.
He was at least as much inspired by scientific interests and his passion for
collecting, also apparent from his numerous botanical and zoological works.
Indeed, the taxonomies used in these branches of biology probably inspired
⁶ See Lamers (2015: 72, 81–2, 166) for the grecità periodization followed here.
⁷ For biographical data see e.g. Fischer (1966); Wellisch (1984: 1–25); Springer and Kinzelbach
(2008: 35–45). On the spelling of his name as ‘Gessner’ see Pyle (2000).
⁸ On Gessner’s linguistic work in general see Colombat and Peters (2009: 17–20).
.  ,     53
Figure 4.3 Conrad Gessner, portrait by Tobias Stimmer (1539–84)

Gessner, ‘the Swiss Pliny’, to engage in a similar endeavour for the diversified
phenomenon of human language (see Ogilvie 2006: 84; Leu and Opitz 2019). His
bio-bibliographical project can likewise be viewed against this background. In his
Universal library (1545–9), he tried to provide a complete overview of Latin,
Greek, and Hebrew authors, their lives, and their works, again alphabetically
organized, in this case by first name. As a result, Gessner has been dubbed both
the ‘father of zoology’ and the ‘father of bibliography’ (Springer and Kinzelbach
2008: 27, 28), to which one might add the title of ‘father of language cataloguing’.
How did Gessner catalogue and classify the speech forms of the world known to
him? And more specifically, what concepts did Gessner employ in order to express
different degrees of linguistic kinship? To answer these questions, I have to take a
look at the very first pages of his Mithridates, where he introduced one of his key
terms: dialectus. The presence of the word in Gessner’s work is not unexpected for
two main reasons. On the one hand, it had been naturalized as a Latin term by the
mid-sixteenth century and Gessner had been using it at least since 1543 (see
Gessner 1543: .6–7). On the other hand, Gessner, trained as a philologist, was
‘especially attached to’ the Greek language and literature, from which the term had
54        
Figure 4.4 Marble bust of the Pontic king Mithridates VI, first century 
Source: Eric Gaba. CC BY-SA 2.5
been borrowed (Leu et al. 2008: 16). He had many Greek works in his library and
was fluent in the language at age fifteen, in 1531, when he played a part in a Greek
performance of a comedy by Aristophanes.⁹ He taught it to fellow students in
Strasbourg, Paris, and Bourges to earn some money in the years 1532–3, and he
was professor of Greek at Lausanne for three years in 1537–40.¹⁰ He moreover
translated many Greek texts into Latin, including works by Porphyry and
Johannes Stobaeus, and edited writings of Aristotle and Marcus Aurelius. He
composed commentaries on Hippocrates and other Greek authors, and he sug-
gested different allegorical interpretations of Homer’s epic poems (Leu et al. 2008:
16–17). Yet his devotion to Greek studies is perhaps most obvious from an early
writing of his, his preface to the 1543 Greek–Latin lexicon he edited, in which he
discussed ‘the utility and pre-eminence of the Greek language in every kind of
studies’.¹¹
⁹ Leu et al. (2008: 16–18), containing an incomplete description of the Greek works in Gessner’s
library.
¹⁰ On Gessner’s knowledge and teaching of Greek see Springer and Kinzelbach (2008: 37, 38, 40).
¹¹ Gessner (1543: .2): ‘   tia Græcæ linguæ, in omni genere studiorum’.
.  !      55
4.3 Ad fontes! Gessner reads Clement of Alexandria
In view of his wide reading in Greek literature, it is no surprise that Gessner’s

initial definition of the word dialectus was entirely taken from an Ancient Greek
work, the Stromata of the Early Christian Greek author Clement of Alexandria.¹²
His Latin rendering of the relevant passage reads as follows in my English
translation:
A dialect, however, is speech exhibiting a peculiar mark or character of a certain

place, or speech showing the proper or common character of a people. In fact, the
Greeks distinguish five differences of dialects in their language, Attic, Ionic,
Doric, Aeolic, and as a fifth the common.¹³
Gessner’s account seems rather dull at first sight and entirely in agreement with
the Greek tradition, a typical instantiation of the humanist call to reach back to
ancient sources: Ad fontes! But the opposite is true, as his rendering of Clement’s
words is not as faithful as it might seem and, much more importantly, it offers a
crucial insight into a key aspect of the metalinguistic apparatus of his Mithridates.
What key aspect do I intend here? In order to answer this question, I need to
compare Gessner’s Latin with Clement’s original Greek at some length. Table 4.1
puts Clement’s original Greek as it was printed in the Florentine editio princeps
of 1550 alongside the first Latin rendering by the productive French translator
Gentien Hervet (1499–1584), printed in the same city one year later, and
Gessner’s Latin version in his Mithridates of 1555. Looking at the table, it
immediately becomes apparent that the Protestant humanist from Zürich did
not rely on the Catholic Hervet’s faithful but somewhat inelegant Latin version,
nor did he translate Clement’s text verbatim. Instead, he repeatedly provided
suggestive renderings and elaborate paraphrases of Clement’s Greek, resulting in
a translation that was considerably longer and much less faithful than Hervet’s.
According to Gessner, the seventy-two original tongues, hai genikaì diálektoi
(αἱ γενικαὶ διάλεκτοι) in Clement’s Greek, should be viewed as languages rather
than dialects. This interpretation is suggested by the fact that he translated the
phrase as ‘dialecti (linguæ potius) communes’, ‘common dialects (languages
rather)’. In what followed, he claimed that there were several dialects subsumed
under each common genus (genus commune), that is to say, under each of the
¹² Clement of Alexandria, Stromata 1.21.142.1–4, quoted in Table 4.1.

¹³ Gessner (1555: 1–2): ‘Est autem dialectus dictio peculiarem alicuius loci notam seu characterem
prę se ferens: uel dictio quæ propriam communémue gentis charactere[m] ostendit. Græci quide[m]
dialectorum suę linguæ differe[n]tias quinq[ue] annotant, Atticam, Ionicam, Dorica[m], Aeolicam, &
quintam communem’. For his interpretation of dialectus see e.g. already Metcalf (2013: 71); Peters
(1970–1: .165–73; 1972: 260, 264–6); Colombat and Peters (2009: 30–2, 34); Stockhammer (2014: 333);
Van Rooy (2016c: 133–4).
56        
Table 4.1 Clement of Alexandria’s Greek compared to Hervet’s and Gessner’s Latin
versions
Clement (1550: 126–7) Hervet in Clement Gessner (1555: 1–2)

(1551: 37)
Εὔϕορος δὲ καὶ ἄλλοι πολλοὶ Euporus aute[m] & alii Euphorus & alij multi
τῶν ἱστορικῶν, καὶ ἔθνη καὶ multi Historici, & gentes & historicorum, gentes &
γλῶσσας πέντε καὶ linguas dicunt septuaginta linguas septuaginta
ἑβδομήκοντα λέγουσιν quinq[ue] cum audiissent quinq[ue] esse dicunt, inde
εἶν[αι], ἐπακούσαντες τῆς vocem Mosis dice[n]tis. nimirum impulsi quòd
ϕωνῆς Μωϋσέως λεγούσης. Erant autem omnes animæ Moses scripsit: Animæ ex
ἦσαν δὲ πᾶσαι αἱ ψυχαὶ ἐξ ex Iacob septuaginta Iacob in uniuersum era[n]t
Ἰακὼβ πέντε καὶ quinq[ue], quæ septuaginta quinq[ue], quæ
ἑβδομήκοντα, αἱ εἰς descenderunt in Aegyptum. in Aegyptum descenderunt.
αἴγυπτον κατελθοῦσαι. Videntur autem ex vera Et sanè uidentur reuera
ϕαίνονται δὲ εἶν[αι] καὶ κατὰ ratione linguæ seu dialecti, dialecti (linguæ potius)
τὸν ἀληθῆ λόγον αἱ γενικαὶ vt vocant, & sermones communes duæ &
διάλεκτοι, δύο καὶ generales esse septuaginta septuaginta, ut in
ἑβδομήκοντα ὡς αἱ ἡμέτεραι duæ, ut nostræ tradunt nostrorum etiam
παραδιδόασι γραϕαί: αἱ δὲ scripturæ. Fiunt autem alię monumentis proditum
ἄλλαι αἱ πολλαὶ ἐπὶ κοινωνίᾳ per communionem duarum reperitur. Reliquæ uerò
διαλέκτων δύο ἢ τριῶν ἢ καὶ vel trium vel etiam plurium multæ sub unum genus
πλειόνων γίνονται· dialectorum. Est autem commune, quod duas aut
διάλεκτος δέ ἐστι λέξις, ἴδιον dialectus, dictio quę loci tres plurésue dialectos
χαρακτῆρα τόπου proprium ostendit contineat, referendę sunt.
ἐμϕαίνουσα: ἢ λέξις ἴδιον ἢ charactere[m], vel, dictio Est autem dialectus dictio
κοινὸν ἔθνους ἐπιϕαίνουσα quæ proprium vel peculiarem alicuius loci
χαρακτῆρα. ϕασὶ δὲ οἱ communem gentis notam seu characterem prę
Ἕλληνες διαλέκτους εἶν[αι] characterem ostendit. se ferens: uel dictio quæ
τὰς παρὰ σϕῖσι, ε ̄ʹ, ἀτθίδα, Dicunt autem Gręci esse propriam communémue
ἰάδα, δωρίδα, αἰολίδα, καὶ quinque apud se dialectos, gentis charactere[m]
πέμπτην τὴν κοινήν: Atticam, Ionicam, Doricam, ostendit. Græci quide[m]
ἀπεριλήπτους δὲ οὔσας τὰς Aeolicam, & quintam dialectorum suę linguæ
βαρβάρων ϕωνὰς, μὴ δὲ co[m]munem, quæ autem differe[n]tias quinq[ue]
διαλέκτους, ἀλλὰ γλώσσας comprehendi nequeunt annotant, Atticam,
λέγεσθαι. voces Barbaroru[m], non Ionicam, Dorica[m],
dici dialectos, sed linguas. Aeolicam, & quintam
communem. Porrò uoces
barbaras (quę scilicet à
Græcis usurpantur) cum
sint incomprehensibiles,
non etiam dialectos, sed
glossas uocari aiunt,
Clemens Alexandrinus libro
1. Stromatéωn.
Note: Underlining marks the phrases most important for my analysis.
seventy-two original common languages. This idea was lacking in Clement’s

Greek original, where, in fact, the emergence of new languages through mixture
was discussed (Van Rooy 2013: 39). Gessner moreover rendered the Greek
.        57
word form dialéktous (διαλέκτους) by means of the very elaborate paraphrase

‘dialectorum suę linguæ differe[n]tias’, ‘differences of dialects of their language’.
The Swiss humanist was clearly struggling with the meaning of the Greek term as
it was used by Clement of Alexandria, which differed from his own understanding
of dialectus. In fact, Clement interpreted the term diálektos as referring, very
generally, to different tongues, presupposing no distinction between language
and dialect. Gessner’s adaptations and paraphrases, however, doubtlessly betray
a conceptual subsumption of dialect under language; indeed, he reserved the term
dialectus principally for forms of speech that resorted under a lingua.
Comparing Gessner’s translation to Hervet’s, the only one available at
Gessner’s time of writing, the reader is at once struck by the fact that the term
dialectus is treated with much less suspicion by the Frenchman. Indeed, unlike
Gessner, Hervet did not read a language/dialect distinction into Clement’s work.
What is more, Hervet elaborately translated hai genikaì diálektoi as ‘linguæ seu
dialecti, vt vocant, & sermones generales’, ‘languages or dialects, as they’—the
Greeks—‘call them, and generic tongues’. This phrasing suggests that Hervet
interpreted dialectus as a Greek synonym for lingua and sermo, even though the
length of his paraphrase indicates that he, too, struggled with Clement’s usage of
diálektos. Anyhow, Gessner’s translation was much more interpretive than
Hervet’s and influenced many later scholars, few of whom were aware that
Gessner was, in fact, quoting Clement and not formulating his own ideas. This
error has persisted even in modern scholarship, in spite of the fact that Gessner
has clearly marked his source at the end of the long quotation.¹⁴ The modern
editors of Gessner’s Mithridates, Bernard Colombat and Manfred Peters, are
exceptional in correctly identifying Clement as Gessner’s source (2009: 31).
4.4 The meanings of dialectus according to Gessner
Following the passage cited from Clement, Gessner elaborated on other meanings
of the word dialectus he had encountered. After observing that the term could also
simply mean ‘articulate speech’, as in Aristotle’s work, or ‘conversation’, as in
Plato’s oeuvre, Gessner described its grammatical sense in his own words as
follows: dialectus was ‘a property of a certain language, either in separate or in
several words, a property by which it differs from the common or from other
similar or cognate [dialects]’.¹⁵ The polysemy upon which Gessner remarked
¹⁴ See e.g. Mylius (1612: 195); Schottel (1663: 150); Schilter (1728: ); Michaeler (1776: preface);
Fischer (1966: 75); Peters (1970–1: .165–6); Ahačič (2008: 96); Simon (2011: 168); Metcalf (2013: 71);
Wilkinson (2016: 211–12).
¹⁵ Gessner (1555: 2): ‘Nos dialectum aliâs simpliciter sermonem siue orationem articulatam
significare obseruauimus, uel ipsum in pluribus uerbis colloquium: aliâs (apud gram[m]aticos
præsertim) linguæ alicuius siue in singulis siue in pluribus uerbis proprietatem, qua à co[m]muni uel
58        
seems to have led to inconsistencies in his own usage of the term dialectus,
especially in relation to lingua (cf. Swiggers 1997: 143). As a matter of fact, lingua
is frequently used to designate varieties he also dubbed dialecti. The Aeolic dialect
of Greek is, for instance, also referred to as Aeolica lingua, ‘Aeolic language’
(Gessner 1555: 5; cf. also fol. 12). Dialectus is moreover found in other, more
general senses throughout his work, such as ‘language’ or ‘way of speaking’ (e.g.
Gessner 1555: 67, 69; cf. Peters 1970–1: .166). In sum, Gessner, as is typical for a
compiler, used the term eclectically, putting the different interpretations he knew
into practice as it suited him and as he found them in his sources. From this
perspective, it is difficult to maintain the somewhat anachronistic statement of
B. Colombat and M. Peters ‘that dialectus very often appears in the Mithridates in
the meaning attributed to it today’.¹⁶ The force of this explanation is moreover
rather limited, as it is unclear what is meant by the modern interpretation of
dialect.
Gessner’s definitions were inspired by the Greek tradition, but his varying usage
of the term was influenced by his eager reading of the works of contemporary
scholars, too. These scholars included, most importantly, theologians such as
his teacher Theodore Bibliander (1504/9–64), who occasionally seems to have
implied that a dialectus historically derived from a lingua, and philologists such
as Johannes Rhellicanus (1477/8–1542), who used the term dialectus to indi-
cate that two related speech forms exhibited superficial linguistic variation
only.¹⁷ In other words, Gessner did not encounter the term dialectus in only
one isolated context but construed a kind of hybrid concept out of the various
usages of the term he came across while reading ancient texts as well as
contemporary scholarship.
How did Gessner understand dialectus as opposed to lingua? His grammatical
definition, quoted above, supplies us with obvious clues. He viewed a dialect as a
specific form of a certain language showing differences both within individual
words and on the level of syntax. Colombat and Peters (2009: 39) add that Gessner
assumed that related dialects also varied slightly on the level of phonetics, phon-
ology, and morphology. This idea might indeed be regarded as implicit in the
examples he provided. One could, however, argue that these aspects are all
covered by Gessner’s explicit but imprecise statement that dialects showed differ-
ences within individual words. For, in making this suggestion, he was no doubt
thinking of the permutatio litterarum, ‘change of letters’, a phrase in which the
reliquis similibus aut cognatis differt’. On this passage cf. Van Rooy (2016c: 133–4), whence the English
translation provided here is adapted.
¹⁶ Colombat and Peters (2009: 32): ‘dialectus apparaît très souvent dans le Mithridate, dans le sens
qu’on lui reconnaît aujourd’hui’. Cf. Stockhammer (2014: 333) for a similarly anachronistic analysis.
¹⁷ For Gessner as Bibliander’s pupil see e.g. Gessner (1555: 78); Amirav and Kirn (2011: ). See
also Chapter 5.
.        59
term littera covered both the letter itself and the sound it represented (cf. also
Chapter 15, Section 15.2). This ancient Roman framework was designed for
etymological purposes by Marcus Terentius Varro (116–27 ) and consisted in
four main letter change operations: deletion (detractio), addition (additio), change
(commutatio, mutatio, permutatio) of a letter, and a position switch (metathesis)
of several letters.¹⁸ Varro’s framework, also outlined by Quintilian, might have
been related to Greek pathology, which emerged around the same time but
comprised much more letter-change processes (see Chapter 2, Section 2.3).
Gessner’s dialect concept remained relatively vague. He nevertheless associated
a dialect and its counterpart language with certain parameters. According to
Gessner’s grammatical definition, a dialect differed not only from the common
language—he no doubt had the Greek Koine in mind—but also from related
dialects. It surely implied that this variation should be superficial, without touch-
ing the core of the language, and close genealogical kinship was clearly a pre-
requisite to label related speech forms dialecti. In fact, elsewhere, Gessner (1555:
21v) repeated the humanist commentator Heinrich Glarean’s (1488–1563) state-
ment that it was justified to give two tongues of ancient Gaul that mutually
differed only slightly one and the same name. Glarean’s (1538: 19) model was
the Greek context, where Aeolic, Attic, Doric, and Ionic were all labelled ‘Greek’
because of their close similarities. Additionally, Gessner claimed that a number of
languages now classified as Baltic, including Lithuanian and Old Prussian, ‘vary
only by dialects’, a formulation suggesting a strong bond between those tongues.¹⁹
In short, Gessner supposed related dialects to differ from one another only
superficially, which may surely be taken to imply that distinct languages displayed
more substantial variation.
It is unclear whether Gessner regarded a dialect as a regional variety of a
language, particular to a certain nation. These two parameters figured in
Clement of Alexandria’s definition of diálektos, in which no subsumption of
dialect under language was presupposed, but they were not explicitly repeated
by Gessner, who usually adopted a very ambiguous attitude towards his sources
(Colombat and Peters 2009: 54). Both parameters might be implicit in his
statement on the Ancient Greek language that ‘everywhere the populace has
something of its own by which it differs from most others’, even though he
principally had social variation in mind here.²⁰ The two parameters were, how-
ever, no doubt presupposed in the classification of the Ancient Greek dialects into
¹⁸ On Varro’s etymological method see e.g. Pfaffel (1981); Taylor (1996: 7–10, passim). On the
permutatio litterarum see e.g. Ax (1987).
¹⁹ See Gessner (1555: 60): ‘ uel Liuorum & Lituanorum & Curorum lingua, eadem est quæ
uetus Prussica. dialectis tantum uariant’. Cf. Gessner (1561: *.4, *.5). On this phrase see also
Chapter 5, Section 5.5.
²⁰ Gessner (1555: 46): ‘habet enim ubiq[ue] uulgus suum aliquid proprium, quibus à cæteris
plerisq[ue] differt’.
60        
Aeolic, Attic, Ionic, and Doric, excluding the Koine, which, unlike the other four,
Gessner did not tie to a Greek tribe or region. Instead, the dialects were subsumed
under the Koine in his view.²¹
Mutual intelligibility does not seem to have been a criterion to speak of related
dialects for Gessner. Lithuanian, for instance, constituted a unitary language
which was divided into four dialects. Speakers of different dialects could not,
however, understand one another very well, Gessner (1555: 59) asserted on the
authority of the Polish geographer Maciej Miechowita (1457–1523), also known as
Matthias de Miechow (Miechowita 1517: E.v). Gessner’s reliance on Matthias’s
description of parts of Eastern Europe is also of interest for another reason. From
Matthias’s work Gessner knew that the Lithuanian linguistic domain—lingua-
gium, the term is Matthias’s—was ‘quadripartite’ (quadripartitum). Matthias, not
yet operating with a clear-cut language/dialect distinction, had described the
speech of the Yotvingians, that of the Lithuanians and Samogitians, that of the
Prussians, and that of the Latvians as one and the same language, even though
there was no mutual intelligibility, except among those having travelled through
the different regions (see Stachowski 2013: 312–13). In an attempt at understand-
ing what Matthias meant exactly, Gessner inserted the following parenthetic
remark in his quotation from the geographer’s work: Matthias ‘seems to mean
that it is one language—he calls it linguagium himself—but distinguished by four
dialects’.²² In other words, Gessner relied on the recently emerged conceptual
distinction in order to understand the views of a scholar to whom the abstract
opposition was apparently not yet known.
Gessner repeatedly associated a dialect with lower social classes, which emerges,
for instance, from his collocation ‘vulgar dialects’ (dialecti vulgares). Dialectus is
contrasted with the lingua communis, ‘common language’, the variety spoken by
learned men (Gessner 1555: 46; 1561: *.4). This phrase clearly derived from the
usual discourse on the Greek context. It is without doubt a loan translation of the
Greek koinē ́ (κοινή), which was short for hē koinē ̀ diálektos, ‘the common way of
speaking’. Gessner, however, found it more adequate to translate this elliptical
designation as lingua communis rather than dialectus communis. He used the
phrase not only when he spoke of the Greek language, but he also applied it to
other languages such as his native German.²³ Gessner also suggested geographical
centrality as a characteristic of a common language. This view implied for him
that the further a variety was removed from the centre, the more it deviated from
²¹ For Gessner’s views on the Greek dialects see Peters (1970–1: .167–72, .188); Colombat (2008: 78–9);
Colombat and Peters (2009: 39–40).
²² Gessner (1555: 59): ‘uidetur sentire linguam unam esse, linguagium ipse uocat, sed quatuor
dialectis distinctam’.
²³ See Gessner (1561: *.4). Cf. Peters (1972: 266). For Gessner’s views on Germanic varieties see
Metcalf (2013: 77–84). For his views on other dialect contexts see Peters (1970–1: .45–191 [passim]).
.  61
the common tongue (Gessner 1555: 58). Finally, in one passage, it seems that
Gessner (1561: *.5) conceived of dialects as ‘descendants’ (propagines) of a
common ancestral language (lingua principalis), thus adding a language-historical
dimension to the conceptual pair, a usage still known today.
Gessner did not evaluate dialectal diversity in general terms as a positive or
negative phenomenon, even though he associated it from time to time with lower
social strata. Yet he attributed certain qualities or deficiencies to individual
dialects. This practice was absent from Roger Bacon’s work, and its presence in
Gessner’s writings might be seen in relation to the ongoing standardization of
vernacular tongues in his days and the competition between dialects that was part
of it. Bavarian German was, for instance, labelled ‘utterly gross’ (crassissima), just
like Doric Greek. Attic was the most elegant Greek dialect, because, Gessner
suggested, it was closest to the common language (1555: 42, 46). However, in
an earlier work, his 1543 praise of the Greek language, he had evaluated the Attic,
Doric, and Ionic dialects all very positively (Gessner 1543: .6–7). The presence
of an imperial court could make a specific dialect elegant, just like the study of
literature, Gessner (1555: 39) argued when discussing the case of Brabantian
Dutch. In short, the phenomenon of dialectal diversity per se was not cheered or
deplored in Gessner’s linguistic work, but individual dialects did receive evaluative
labels, both negative and positive.
4.5 Conclusion
‘I make Swiss a dialect of German’. Conrad Gessner actively subsumed certain

speech forms as dialects under specific languages, and he did so independently
from Roger Bacon, whose linguistic work was largely unknown in the sixteenth
century. Gessner’s relating of linguistic entities in this manner should be framed
within his broader intellectual programme, as he tried to throw the enormous
amounts of information with which he and his colleagues were confronted into
sharper relief. In his catalogues, he created rudimentary groupings of natural
objects, involving not only the vast variety of animals and plants of the world
but also the great diversity of human language. In the latter case, he made use of a
distinction between language and dialect to do so. Even though Gessner did not
define the conceptual pair very clearly, a close analysis of the definitions in his
work and his terminological usage has made it possible to identify the parameters
he associated with both concepts.
Unlike Bacon, Gessner was not an isolated case. He was part of a wider intellec-
tual trend and may even be seen as an early culmination of it. In order to understand
his position in this complex development, I should zoom out in the next few
chapters, which trace the sixteenth-century genesis of the language/dialect pair
from a bird’s-eye perspective, while framing it in its intellectual context.
5
Lingua and dialectus
From synonymy to contrast
Conrad Gessner was an eager early adopter of the distinction between language
and dialect, but he was most certainly not the first humanist to assume its
existence and put it to use in his linguistic work. Instead, he followed in the tracks
of scholars at their acme in the first half of the sixteenth century. Where exactly do
we perceive the first traces of an assumed distinction between the two concepts?
And under what circumstances did this occur?
5.1 Dialectus as a Latin word
When in the 1470s the Italian humanist Niccolò Perotti (1429/30–80) was draw-
ing up his Cornucopia, an impressive commentary on Martial turned encyclopae-
dia, the fact that certain languages show regional differences was widely
acknowledged. In his posthumously published bestseller, Perotti pointed out
that there was ‘sometimes diversity of speech within one language’, which he
exemplified, not surprisingly, by means of the Greek tongue. What is remarkable,
however, is Perotti’s remark that just as an individual language is, such diversity of
tongue ‘is likewise called lingua’, a polysemy he did not problematize.¹ Perotti was,
in other words, not yet operating with a concept of dialect clearly distinct from,
and subsumed under, language. He did, most likely, not yet have a notion of a
linguistic subsystem within a language. His assumption that variation within
a language was not a universal phenomenon perhaps kept him from positing
such a general distinction. Perotti stated, after all, that there was ‘sometimes
diversity of speech’, a view doubtlessly informed by the widespread Renaissance
idea that the Latin language was exempt from dialectal diversity (see Van Rooy
2018a on this idea). In the eyes of Perotti, who devoted most of his attention to
Latin and accorded a merely auxiliary position to Greek (Charlet 2013), this lack
of variation was obviously a positive property. As he was trained as a Hellenist,
¹ Perotti (1489: 85): ‘Quin etiam in vna lingua est aliquando sermonis diuersitas: quæ similiter
lingua dicitur: ut in lingua græca: est lingua: quæ co[m]munis vocatur: & Attica: & Dorica: & Ionica: &
Edica [sic]’. Cf. Trovato (1984: 208–9).
DOI: 10.1093/oso/9780198845713.001.0001
.      63
Perotti must have been familiar with the Greek term diálektos. Its Latin form
dialectus was for him, however, not an approved word, which is surely why he did
not rely on it when describing diversity within one language, even though the
Greek language with its dialects was the prototypical example for this phenom-
enon. So what is the background of Latinized dialectus? And what is its relevance
for the genesis of the conceptual pair?
The late fifteenth-century Latinization of the word and its subsequent natural-
ization as a Latin word were largely due to the false assumption of many human-
ists that the Roman rhetorician Quintilian had already borrowed Greek diálektos
into Latin as dialectus in a much-read passage (Van Rooy 2019). For fifteenth- and
early sixteenth-century editors of Quintilian’s Institutes of oratory preferred at
1.5.29 the variant reading dialectos, the accusative plural of Latinized dialectus,
which is extremely rare in the manuscript tradition, above the more common
reading dialectus. The latter form is the transcription of the Greek accusative
plural dialéktous (διαλέκτους) into the Latin alphabet and is accepted in most
modern editions of Quintilian’s text, because of its prevalence in the manuscript
tradition. As a result of the preference of humanist editors, Latinized dialectus was
mistakenly taken to be an ancient term by many humanists. Soon it was natur-
alized as a Neo-Latin word, even though most humanists did not regard it as a new
constituent of the Latin lexicon. It quickly spread across the Republic of Letters,
from Italy to Switzerland and Germany and thence to the rest of Europe.
The borrowing of the term dialectus into Latin, the scientific lingua franca of
the Renaissance, and its subsequent naturalization constituted a key intermediary
step in the genesis of the conceptual pair. What is more, the fact that early
sixteenth-century scholars started to contrast lingua to dialectus is the foremost
symptom of the emergence of a language/dialect distinction, as I will argue in the
next section. Yet before moving on to the contrastive usage of the terms lingua and
dialectus, it may be useful to emphasize that fifteenth- and early sixteenth-century
glosses and definitions of the word dialectus and its Greek original diálektos do not
reveal a straightforward distinction between language and dialect. It is variously
interpreted as idioma loquendi, ‘particularity of speaking’, proprietas linguae or
loquendi, ‘property of tongue’ or ‘of speaking’, and genus loquendi, ‘kind of
speaking’.² Arguably, one could interpret the phrase proprietas linguae in terms
of a subsumption of ‘dialect’ under ‘language’. It seems, however, more plausible
́
that this reflects the traditional Greek definition of diálektos as glōssēs idíōma,
and, as I have suggested earlier (Chapter 2, Section 2. 2), this phrase should be
understood as ‘property of tongue’, as a particular linguistic form different from
others, and not as a linguistic entity intrinsically subsumed under a language.
² See e.g. Crastone ([1478]: s.v. διάλεκτος), Sabellicus (1490: 64), and Calepino (1502: s.v. dialectus),
respectively.
64   
The very frequent synonymic use of dialectus and lingua in the early Renaissance
seems to confirm my analysis. In a late fifteenth-century commentary on
Suetonius, for instance, one reads that ‘the Rhodians speak in the Doric dialectus
and lingua’.³ It is likely that this synonymy of lingua and dialectus was borrowed
from the Greek language, in which glôssa and diálektos were usually used inter-
changeably, too (Lambert 2009: 21).
To conclude, in the late Quattrocento and early Cinquecento, there was, much
as it had been in antiquity and the Middle Ages, a direct awareness of the fact that
individual languages could manifest themselves in different forms, which was
expressed by often elaborate circumlocutions. Aeolic, for instance, was regarded as
one manifestation of the Greek language. No early Renaissance scholar operated,
however, with more abstract concepts categorizing one variety, a dialect, as clearly
subsumed under another, a language. In other words, a language was not yet seen
as an entity roofing different dialects, but different varieties were regarded as
constituting different manifestations of one language—contrast Figure 5.1 with
Figure 5.2 in this regard. In the first half of the Cinquecento, a change occurred,
and the terms lingua and dialectus were increasingly contrasted in several different
ways, which are all related and which all appear in the thought of both Conrad
Gessner and several of his immediate predecessors and contemporaries.
manifestation a manifestation b manifestation c manifestation d manifestation e

(e.g. Aeolic) (e.g. Attic) (e.g. Koine) (e.g. Doric) (e.g. Ionic)
L A N G U A G E
Figure 5.1 Conceptualization of a language (e.g. Greek) before the early Cinquecento
language
(e.g. Greek [Koine])
dialect a dialect b dialect c dialect d

(e.g. Aeolic) (e.g. Attic) (e.g. Doric) (e.g. Ionic)
Figure 5.2 Conceptualization of a language (e.g. Greek) emerging in the early

Cinquecento
5.2 The key symptom: contrasting lingua to dialectus
As Conrad Gessner tried to understand Clement of Alexandria’s comment on the

seventy-two original tongues of the world, hai genikaì diálektoi in Greek, the Swiss
³ Beroaldo (1493: 138): ‘Rhodii Dorica dialecto et lingua loquuntur.’

.   :     65
humanist intuitively sensed that, in his view, Clement had not used an adequate
word to refer to these languages. Although rendering Clement’s diálektoi initially
as dialecti, Gessner felt the need to add between parentheses ‘languages rather’
(‘linguæ potius’). This addition indicates that for the Swiss humanist, a lingua was
something different from a dialectus. Other sixteenth-century scholars, too,
opposed the terms lingua and dialectus, and even more explicitly and emphatically
so. Such contrastive statements were nearly absent from earlier thought, and their
large-scale appearance constitutes the single most compelling piece of evidence for
the early sixteenth-century emergence of the conceptual pair. Consider two
prominent examples.
The Rotterdam humanist Desiderius Erasmus (c. 1466/7/9–1536) provided a
very early instance when explaining the Greek term diálektos in the 1519 edition
of his commentary on Acts (cf. Figure 5.3):
To the Greeks, dialect is a property or fashion of tongue. For example, among the
Greeks, even though there is one language (lingua), there are nevertheless five
Figure 5.3 Desiderius Erasmus holding a Greek book, by Hans Holbein, 1523
Source: Wikimedia Commons. Public domain
66   
dialects (dialecti), so that he who is versed in Greek can directly recognize

whether a speaker is Attic or Doric, Ionic or Lacedaemonian.⁴
Erasmus was contrasting lingua to dialectus here, even though one could argue
that the Dutch humanist still imagined the five dialects to be manifestations of the
Greek language rather than entities subsumed under it. It is nevertheless not
unlikely that he equated the Greek language with the Koine, under which he
placed the five Greek dialects Aeolic, Attic, Doric, Ionic, and Lacedaemonian.
Here, Erasmus did not mention Aeolic, but it emerges from his other works that
he was well aware of its existence (e.g. Erasmus 1528: 169–70).
There is a clearer example still in the work of the Spanish humanist pedagogue
Juan Luis Vives (1492/3–1540), writing in the Southern Low Countries some
fifteen years after Erasmus:
There are some who introduce characters speaking each not only in their own
language (lingua), but also in their own dialect (dialecto), as the sophist Lucian does
in his Sale of creeds. This would be easy to do in those languages that have various
dialects, as the Greek formerly had, and all vernaculars now per nation. In Latin,
this could hardly happen, unless perhaps by means of the distinction in periods.⁵
The Spanish humanist’s contrasting of lingua and dialectus clearly concerned

different types of linguistic entities. Different languages were distinct speech
forms, whereas cognate dialects resorted under one and the same language, for
which Greek and the vernaculars of Western Europe served as examples. Latin,
however, was not regionally diversified, even though it could be distinguished into
different periods. Here, Vives was thinking of the traditional division of the
history of the Latin language into four periods: ancient, Latin, Roman, and
mixed (e.g. Denecker 2017: 229–32).
The contrastive usage of the terms lingua and dialectus did not emerge over-
night. The late fifteenth to the early sixteenth century was the transitory period in
which this gradually became a widely accepted given among humanist scholars,
who increasingly encountered passages in which the two terms were contrasted or,
in some instances, only seemingly so. A case in point is the Bolognese Hellenist
Antonio Urceo (1446–1500), also known as Codro, who in 1485/6 briefly
remarked of Greek that ‘it is difficult and hard to learn a foreign language with
⁴ Erasmus (1519: 202): ‘Græcis dialectus est linguæ proprietas aut species, uelut apud Græcos cum
una sit lingua, quinq[ue] tame[n] sunt dialecti, ut qui Græce calleat mox possit agnoscere, Atticus sit
qui loquatur an Doricus, Ionicus an Lacedæmonius’.
⁵ Vives (1533: .vii): ‘Sunt qui personas loquentes introducunt sua que[m]q[ue] non lingua modo,
sed etiam dialecto, vt Lucianus sophista in auctione uitarum, quod pro[m]ptum esset facere in ijs
linguis, quę dialectos habent varias, vt fuit olim Græca, & nunc vulgares omnes nationatim. In Latina
ægre fieret, nisi forte ex distinctione ætatum’.
.       67
five dialects’.⁶ Here, a contrast between the two terms seems to have been
presupposed. Several years later, however, probably in 1491/2, Codro explicitly
treated lingua and dialectus as synonyms: ‘for there are five dialects or languages
of the Greeks’.⁷ As Kozma Ahačič (2008: 97) has rightly pointed out, such usage
could be dubbed a ‘terminological inconsistency’ from a modern point of view,
even though it is more fruitful to understand it as a ‘terminological reality’, which
must be taken into account in writing a history of the conceptual constructs
underlying it.
Codro’s case demonstrates that it can be misleading to rely on a single passage
to claim that a language/dialect distinction was presupposed by its author, let
alone his contemporaries. Even so, the conclusion remains that a number of
leading humanists started to use the terms dialectus and lingua contrastively in
the first decades of the sixteenth century. They may have been inspired to do so by
their reading of assertions such as Codro’s first observation. It might be useful to
borrow the concept of bridging context from linguistics to clarify the impact
expressions like Codro’s may have had. A bridging context ‘allows for two
interpretations of the linguistic item, the original interpretation and the new,
pragmatically inferred interpretation’ (Gipper 2014: 87). In other words, a lin-
guistic item x with meaning  can receive a new meaning  in specific contexts.
After a while, meaning  can come to overshadow meaning  and even make it
obsolete. Applying the concept of bridging context to dialectus, I can say that
dialectus at first only had the generic meaning of ‘way of speaking; property of
tongue’, but that it was sometimes used in such a way by certain scholars,
including Codro, that another interpretive possibility came about: dialectus as
‘variety subsumed under a language’. A number of scholars reading Codro’s work
would, by consequence, have been induced by the humanist’s usage of the term to
read a contrastive interpretation into it. This originally context-bound reading
resulted in a new meaning for dialectus, which did, however, coexist for a long
time with the original, more generic meaning of the word.
The direct and explicit opposing of lingua to dialectus is the key symptom of the
emergence of the conceptual pair, but the subsumption of dialect under language
can be recognized in several other ways, too, all related to this contrastive usage.
5.3 Developing the concept of common language
In the first half of the sixteenth century, humanists developed the concept of
common language, usually termed lingua communis and perceived as roofing
⁶ Codro (1502: .ii): ‘Atqui difficile et arduum est peregrinam linguam quinq[ue] dialectos
habe[n]tem ediscere’.
⁷ Codro (1502: .v): ‘Su[n]t enim græcorum dialecti seu linguæ quinq[ue]’.
68   
related dialects in the way suggested in Figure 5.2. I have illustrated how Conrad
Gessner characterized a dialectus as a variety of speech differing from related
dialects as well as from the lingua communis, a kind of supraregional variety used
by educated speakers to communicate with one another.⁸ Whence did this concept
originate?
The new knowledge of the Greek language played a key role in this evolution.
On the one hand, the Latin phrase lingua communis was inspired by the Greek
elliptical designation hē koinē ̀ [diálektos], ‘the common [way of speaking]’. In
Latin, substantivizing adjectives was more difficult because of its lack of an article;
as a result, the noun lingua was added to communis to render the Greek phrase in
Latin. At first, the Greek Koine was widely conceived as one of the five main
manifestations of the Greek language. Gradually, however, it came to occupy a
place clearly distinct from the other dialects in sixteenth-century grammatical
works on Greek. The humanist Philipp Melanchthon (1497–1560), for instance,
clearly did so in his popular grammar of the language, first published in 1518
(see Melanchthon 1518: a.i).
On the other hand, the concept of common language was extracted from the
Greek context and soon applied to vernacular tongues, despite the fact that many
of them still lacked a widely accepted norm. On the Italian peninsula, for instance,
Vincenzo Colli (c. 1460–1508), a poet also known as ‘Calmeta’, proposed as his
answer to the questione della lingua, with explicit reference to the Greek context, a
mixed common language based on various Romance varieties spoken at the
Roman court. He did this in his Della volgar poesia, a book on vernacular poetry
now lost to the ages, but we are well-informed about Calmeta’s ideas through
Pietro Bembo’s (1470–1547) refutation of them.⁹ Examples from the Italian
context could easily be multiplied for the early sixteenth century. In the second
half of the Cinquecento, however, the design of such a common Italian variety was
judged impossible by scholars such as Benedetto Varchi (1503–65) and Vincenzo
Borghini (1515–80).¹⁰ For German-speaking territory, an early example can be
found in Melanchthon’s Greek grammar:
The speech common to all is called common language, just like, with us, there is a
certain fashion of speaking common to the Swabians, Bavarians, and Ubians.
Each of them nevertheless has its own particularities.¹¹
⁸ Chapter 4, Section 4.4. On the insufficiently studied history of the common language concept see
Regis (2012), with specific reference to the Italian context.
⁹ Bembo (1525: –). On Calmeta’s views see e.g. Pozzi (1978: 92–3); Trovato (1984: 216–17).
¹⁰ See Varchi (1570: 269–71); Borghini (1971 [before 1580]: 338, 341–3).
¹¹ Melanchthon (1518: a.i): ‘Qui sermo co<m>munis omnibus est, lingua co<m>munis dicitur
perinde ut apud nos est aliqua ratio loquendi communis Sueuis, Boiis, Vbiis, singulis tamen sui sunt
idiotismi’.
.       69
An intuitive comparison of the Greek and his native German contexts led
Melanchthon to oppose lingua communis to dialecti. It is, however, not yet
entirely clear whether he interpreted the linguistic particularities of the
Swabians, Bavarians, and Ubians as independent entities subsumed under a
common language. He suggested rather that these segments of the German-
speaking population had a common speech that showed an occasional linguistic
particularity, without presupposing two separate linguistic levels. This analysis
can, however, not be maintained for his views on the Greek context, as he clearly
set the lingua communis apart from the Attic, Ionic, Aeolic, and Doric dialects
subsumed under it.
Melanchthon’s Protestant partner-in-arms Martin Luther (1483–1546) had a
more specific linguistic entity in mind when talking about the German lingua
communissima he adopted in his writings. In one of his table talks held between
1530 and 1545, Luther used this phrase, marked by the superlative ‘most com-
mon’, to characterize the German of the Saxon chancellery. This variety, he
clarified, enabled him to communicate with the entire German-speaking area. It
was, after all, employed by ‘all princes of Germany’ as the language of adminis-
tration.¹² The wide usage of this variety was probably exaggerated by the reformer,
but Luther’s activities did, in fact, lead to this form of German becoming the main
fundament of the later German standard language.
The concept of common language was, remarkably enough, first applied to
linguistic areas that lacked a centralized government to speak of and where
language standardization was not as straightforward as in more centralized states:
the Italian peninsula and German-speaking territories (cf. Burke 2004: 109).
Precisely this lack of political unity might have stimulated scholars of these
regions, some of them not devoid of patriotic sentiments, to conjure up a common
language by way of compensation. It no doubt also relates to the fact that Greek
studies boomed early in those areas. Soon the concept of common language was
extrapolated to vernacular languages other than German and Italian, too, most
notably French, Dutch, Spanish, and English (see e.g. Zimont 2020 for French).
Furthermore, already in 1531, Juan Luis Vives (1531: 79) promoted, in general
terms, a common form of speech as a welcome instrument to overcome commu-
nication problems caused by dialectal diversity and to create a societal bond.
In summary, one could say that a double evolution occurred. On the one hand,
the Greek Koine came to be sensed by a large number of humanists as something
fundamentally different from the other dialects of Greek they knew. On the other
hand, scholars abstracted the concept of common language from the Greek
context in the course of the early sixteenth century and gave it a wider application.
¹² Luther (1913: 639): ‘Ich rede nach der Sechsischen cantzlei, quam mutuantur omnes principes
Germaniae’. For lingua communissima see Luther (1913: 640); Burke (2004: 101). Cf. Luther (1916:
78–9); Borst (1957–63: 1067–8).
70   
This is, however, too simplified an account, since the relationship between Greek
and the Western European vernacular languages was not simply that of a model
being applied to new contexts. As a matter of fact, scholars often relied upon their
contemporary linguistic context to understand the Greek state of affairs, as, for
instance, Petrus Antesignanus (1525–61) did in his commentary on the Greek
grammar of the Brabantian humanist Nicolaus Clenardus (c. 1493/5–1542).
Antesignanus, a Protestant French Hellenist who was also active as a Hebraist
and musician, offered a window on contemporary French dialectal diversity,
which he discussed for the student’s benefit:
In order that the entire fact of this matter is put before your eyes and you fully
understand what the grammarians mean by particularities (idiomata), I will
uncover the whole affair for you with few words by means of the example of
our French speech.¹³
In what followed, Antesignanus stated that almost all people wrote and spoke
French in the area subjected to the French throne. Nevertheless, not everyone
talked as neatly as the people at the royal court and in Paris. This fact made it easy
to judge from which region a speaker originated, despite their efforts to avoid
recognition. There were some exceptions, however. Some succeeded in forgetting
their native dialect and in speaking neatly and purely the common French tongue.
They even managed to enrich their French with some phrases and sayings derived
from the dialects of the famous cities that were praised for their language. Dialects
that had received general recognition could, in other words, be used as flowers
adorning the common language (Antesignanus 1554: 11–12).
The French humanist then proceeded to the Greek context. He claimed rather
anachronistically that all Greeks tried to speak and write the Koine, the precepts of
which were transmitted through grammatical writings. Everyone nevertheless
retained particularities from their native dialect, as did, for example,
Hippocrates, whose speech exhibited features from his native Ionic dialect.
Antesignanus also mentioned Attic properties as the features par excellence for
the enrichment of the Greek common language.
Antesignanus thus clearly suggested parallels between French and Greek diver-
sity. What is more, he used his native French context to clarify the Greek state of
affairs, which entailed the retro-projection of a contemporary and familiar onto a
past and unfamiliar context (cf. Van Rooy 2020a: Chapter 8). At the same time,
however, his views on the Greek dialects were likewise partly determining his
¹³ Antesignanus (1554: 11): ‘Vt autem totum huiusce rei negotium tibi ob oculos proponatur, ac
quid Grammatici per idiomata intelligant, penitus noscas, exemplo nostri Gallici sermonis tibi paucis
rem omnem aperiam’ (translation adapted from Van Rooy 2016c: 131). What follows is mainly based
on Van Rooy (2016c: especially 131–3).
.       71
views on French diversity. After all, his idiosyncratic insistence on the idea that
French had approved dialects that could lexically adorn the common variety
appears to have been partly inspired by the existence of literary dialects in
Greek. Figure 5.4 offers a schematic visualization of the way in which
Antesignanus coordinated the French and Greek contexts with each other.
conscious clarification
Greek prototype French diversity
unconscious fashioning
Figure 5.4 Antesignanus’s perception of the relationship between Greek and French
variation
Source: adapted from Van Rooy (2016c: 132)
Antesignanus’s intuitive comparison of French and Greek diversity should be seen

in terms of a process of mutual reinforcement. These two dialect contexts were indeed
similar to a certain degree, but Antesignanus suppressed their differences in order to
stress their supposed similarities. The alleged basis for comparability was the idea that
both Greek and French diversity were characterized by a common language and
some generally approved dialects, by which the common variety could be enriched.
Some renowned compatriots of Antesignanus, including Pierre de Ronsard
(1524–85), made a similar suggestion, even though French dialects were never as
prestigious as the canonical literary varieties of Greek (e.g. Ronsard 1565: 4–5).
In sum, Antesignanus’s case suggests that the reason why the Greek Koine was
increasingly regarded as an entity clearly distinct from the other Greek dialects
cannot be exclusively sought in the Renaissance appropriation of Greek scholar-
ship. It was at least as much fostered by the sociolinguistic reality of sixteenth-
century Europe. It seems, in other words, highly likely that even though most
scholars explicitly mentioned the Greek Koine as their model common language—
Luther being a notable exception—the existence of vernacular supraregional
koines and emerging standard languages reinforced the tendency to set the
Greek common language apart from the other Greek dialects. It is therefore safe
to state that the concept of common language emerged as the result of accom-
modating the Greek and vernacular linguistic contexts to each other. This process
involved two important steps: first, the intuitive comparison of the Ancient Greek
language with a more familiar vernacular state of affairs, often with a model status
for the Greek Koine; secondly, the abstraction of the concept of common language
from the situations compared, often involving a suppression of the differences
between them. As a result, it was now possible for scholars to conceptually
superpose a common language on its related dialects. The common language
72   
was closely associated and eventually identified with the linguistic norm, from
which dialects deviated, and can, in this capacity, be viewed as a forerunner of the
later concept of standard language.
5.4 Updating dialectus definitions
I have shown that the definition Conrad Gessner (1555: 2) generically attributed
to the grammarians explained dialectus as ‘a property of a certain language
(proprietas linguæ alicuius), either in separate or in several words, a property by
which it differs from the common or from other similar or cognate [dialects]’ (see
Chapter 4, Section 4.4). Such further elaborations of dialectus definitions serve as
yet another indication of the large-scale emergence of the conceptual pair in the
course of the early sixteenth century. The addition of determiners such as alicuius
(‘certain’) and unius et eiusdem (‘one and the same’) to widespread definitions à la
proprietas linguae suggest that its authors understood dialectus as a ‘variety of a
specific language’ rather than as a ‘property of tongue’ in generic terms. An early
example is the definition of the humanist rhetorician Petrus Mosellanus
(1493–1524), who defined the Greek word diálektos, admittedly still vaguely, as
‘a certain property, preserving the peculiar character of any particular language’ in
his commentary on Quintilian’s Institutes of oratory.¹⁴ According to Mosellanus, a
diálektos preserved the core characteristics of a language but had some particu-
larity proper to it and distinguishing it from the language under which it resorted.
He thus seems to have been reading a subsumption of dialect under language into
the definition of diálektos formulated by the late Byzantine scholar Gregory of
Corinth, which he was translating rather suggestively and interpretatively
(cf. Chapter 2, Section 2.2, for Gregory’s definition).
Mosellanus and Gessner were certainly not alone, and many additional
instances could be cited, but let me limit myself to revealing examples from
sixteenth-century England, where several prominent lexicographers expanded
on traditional definitions of dialectus. Sir Thomas Eliot (c. 1490–1546), for
instance, was the first Englishman to define Latin dialectus in his successful
Latin–English dictionary, first published in 1538. Modelled on Calepino’s Latin
lexicon, Eliot’s Dictionary explained the term as ‘a maner of speche, as we wolde
saye diuersities in englysshe, as Northerne speche, Southerne, Kentyshe,
Deuenishe, and other lyke’ (1538: ). Interestingly, he applied dialectus to
his native context, at the same time opposing the level of the English language to
that of its varieties. In another Latin–English dictionary, that of Thomas Cooper
(1565: s.v. dialectus), a shorter but related explanation was offered: ‘A maner of
¹⁴ Mosellanus (1527: 42): ‘proprietas quædam, peculiare[m] alicuius linguæ particularis

charactere[m] serua[n]s’.
.    73
speache in any language diuers fro[m] other’. Again, we see that the level of maner
of speache was opposed to that of language but this time without exemplification.
In a number of English–Latin dictionaries, similar definitions can be found.
Richard Howlet (fl. 1552) picked up Eliot’s term diversitie and rendered the
English phrase ‘Diuersitie of speache wythin one realme’ in Latin as dialectus.
Howlet (1552: .iiii.) seems, in other words, to have interpreted dialectus as a
regional variety of the language of a realm, much like Eliot before him, to whom he
was indebted. It is, however, in a late sixteenth-century English–Latin lexicon that
the two conceptual levels came to the fore most obviously. The Churchman John
Rider (1562–1632) believed dialectus to be the Latin equivalent of the lengthy
English phrase ‘A propertie of speech, diuerse from the rest of the same language’
(1589: 460). A language that was otherwise the same enclosed within itself several
varieties that differed from one another.
Looking outside the lexicographical tradition, I can remark that dialect was
gradually naturalized as a valid English word from the late 1560s onwards. Early
adopters of the word often still felt the need to gloss it in their works, suggesting
that its acceptance as an approved English word took some time. Richard
Stanihurst (1547–1618), for instance, discussed language diversity in his
Description of Irelande in correlation to its geography:
As the whole realme of Ireland is sundred into foure principall parts, as before is
sayd, so eche parcell differeth very much in [the] Irishe tongue, euery country
hauing his dialect or peculiar maner, in speaking the language.
(Stanihurst 1577: 4)
Here, too, two distinct levels of linguistic entities were presupposed: the Irish
language and the different ways in which it was spoken, its dialects. As in Howlet’s
dictionary, the level of language was tied to a ‘realm’ by Stanihurst.
In parallel with the expanding and updating of definitions, circumlocutions
such as dialect(s) of language x began to appear with increasing frequency. These
phrases no doubt implied a subsumption of dialect under language. It is, for
instance, very obvious in Conrad Gessner’s statement that he ‘make[s] Swiss a
dialect of German’, discussed in the previous chapter. Such phrases appeared
most often when scholars made reference to the vernacular or Greek dialects,
but they occasionally also surfaced in the context of the so-called Oriental
tongues. For instance, the great similarity Syriac and Punic exhibited to
Hebrew led the Italian humanist Angelo Canini (1521–57) to conclude ‘that
they have to be called dialects of it’.¹⁵ For this insight, Canini was very much
¹⁵ Canini (1555: ã.4): ‘vt eius Dialecti nominari debeant’.

74   
inspired by the Greek language and its dialects, which were the main subject of
his book, entitled Hellenism.
5.5 The coining of a new phrase: ‘to differ only in dialect’
When Conrad Gessner commented on the languages now known as Baltic, he

stated that the Livonians, Lithuanians, and Curonians had the same language as
the ancient Prussians, to which he added that they ‘only vary in dialects’.¹⁶ In
doing so, Gessner intended to emphasize the close relationship between these
tongues, which were claimed to exhibit only superficial differences, namely on the
level of dialect and—it seems implied—not on the level of language. Gessner did
not coin this Neo-Latin collocation, which was, however, still very recent at his
time of writing. How did it come into being and in what context?¹⁷
The phrase ‘to vary only in dialect(s)’ is likely to owe part of its origins to two
widely used and frequently reprinted Renaissance grammars of Greek, composed
in the first decades of the sixteenth century by the German-speaking authors
Philipp Melanchthon and Jakob Ceporin (c. 1499/1500–25). In their grammars,
certain parts of speech were said to ‘vary in dialects’ (variare dialectis). Ceporin,
for instance, described in a paragraph entitled ‘Variation of dialects’ ‘how the
prepositive and subjunctive articles vary in dialects’.¹⁸ The prepositive article
corresponds to the Greek article ho, hē, tó (ὁ, ἡ, τό), whereas the so-called
subjunctive article refers to the relative pronoun hós, hē ,́ hó (ὅς, ἥ, ὅ). In the
paragraph on these article forms, Ceporin related how ‘the Dorians turn ē into ā
also in the articles: hā timā ́ [‘the honour’]’.¹⁹ The interpretation of dialectal
differences as accidental and superficial variation was not yet explicitly present
in the phrase variare dialectis, even though later scholars may have understood it
in this manner in bridging contexts. In other words, certain applications of the
phrase variare dialectis, ‘to vary in dialects’, evoked a new, more specific inter-
pretation: ‘to vary only in dialects’.
Superficial variation was most certainly implied when in 1536 the South-
German Protestant pastor Andreas Althamer (c. 1500–c. 1539) stated that he
considered two German words to be ‘varying somewhat in the German dialect’.²⁰
The earliest printed testimony of the phrase ‘to differ only in dialect’ identified
thus far, however, dates to 1543, when it appeared in a posthumously published
¹⁶ Gessner (1555: 60): ‘dialectis tantum uariant’.

¹⁷ See Van Hal and Van Rooy (2017) for a more elaborate discussion of the genesis and uses of this
phrase.
¹⁸ Ceporin (1522: .v): ‘V ’; ‘Vt articulus præpositiuus & subiu[nctiuus]
dialectis uaria[n]t’. Cf. Melanchthon (1518: p.iii).
¹⁹ Ceporin (1522: .v): ‘Dores uertu[n]t α in η [sic pro η in α] etia[m] in articulis. ἁ τιμά’.
²⁰ Althamer (1536: 275): ‘uariante nonnihil germanica dialecto’.
.      : ‘    ’ 75
commentary on Caesar’s Commentaries on the Gallic war, composed by the

philologist and Neo-Latin poet Johannes Rhellicanus (1477/8–1542), born in
Rellikon near Zürich:
Hence, I do not doubt that Polybius, Pliny, Tacitus, and other writers who visited
Germany and Gaul would have indicated that they had the same language, if they
varied only in dialect one from another.²¹
Briefly, the phrase seems to have been transposed from early modern gramma-
ticographic discourse on Greek to other textual genres such as language catalogues
and especially commentaries on ancient historiographers in which the relation-
ship between ancient Gauls and Germans played a prominent role.²² Significantly,
this transposition involved the addition of the specification ‘only’ to the phrase ‘to
vary in dialect(s)’. This development suggests that the term dialectus was under-
going a semantic specialization towards a more concrete meaning: ‘variety of a
language differing only superficially from other varieties of the same language’,
instead of the vaguer interpretations of ‘way of speaking’ and ‘property of tongue’.
From the 1560s onwards, the Latin phrase appeared with increasing frequency.
Its most common terminological guises appear to have been dialecto tantum
variare or sola dialecto differre, with only the word dialectus as a fixed element
in the expression. The influential historian and political philosopher Jean
Bodin (1529/30–96), for instance, propounded in his 1566 Method for the easy
comprehension of history that a wide range of Slavic peoples ‘use the same
language . . . and differ only in dialect’, thus giving a very wide interpretation to
the term lingua, which transcended political borders.²³
The Neo-Latin collocation, which served different purposes, proved to be very
popular and was soon borrowed into several European vernaculars. It was adopted
remarkably early in English. In 1566, the Jesuit John Rastell (1530/2–1577) had
already stated that, in the Slavic tongue, ‘the difference consiste in Dialecte and
proprietie only’, in which he may have been inspired by Bodin’s work, printed
some months previously, even though he referred only to Gessner’s language
catalogue in the margin (see Rastell 1566: 65, misnumbered as 75). A few pages
further, several Germanic varieties were said to ‘differ but in Dialect’, even though
they were not mutually intelligible (Rastell 1566: 68). Rastell was, however, not
only the first author to use the phrase in English, but he also provided the first
²¹ Rhellicanus (1543: 73): ‘Proinde mihi dubium non est, quin Polybius, Plinius, Tacitus, & alij
scriptores, qui Germaniam, & Galliam inuiserunt, indicassent eandem utrisq[ue] linguam fuisse: si
dialecto tantum inter se uariassent’.
²² See Van Hal (2013/14) for humanist views on the relationship between Celtic and Germanic.
²³ Bodin (1566: 439): ‘sic enim audio Polonos, Bohemos, Rußios, Lithuanos, Moschouitas, Boßinios,
Bulgaros, Seruios, Croatios, Dalmatas, Vandalos eade[m] Sclauorum vti lingua, quæ in Sca[n]dia
vsurpatur, ac sola dialecto differre.’
76   
printed instance of the English term dialect found thus far (Van Rooy and
Considine 2016: 647–51).
Rastell’s testimony suggests that the phrase ‘to differ only in dialect(s)’ origin-
ally meant something like ‘to differ only in particularity of tongue’. It may in other
words have been inspired by a very traditional definition of dialectus, which did
not presuppose a subsumption relation between dialect and language: proprietas
linguae, ‘particularity of tongue’. The fact, however, that dialectus appeared very
often in the vicinity of terms meaning ‘to differ’ and especially ‘only’ seems to have
stimulated the contrasting of dialect with language. What is more, as I will argue
in Chapter 8, it became customary towards the end of the sixteenth century to
extend the phrase by introducing the ‘language’ pole into it, resulting in colloca-
tions such as ‘to differ not in the entire language but in dialect only’.
In short, the widespread idea that two varieties ‘differed only in dialect’ not only
presupposed but also reinforced an opposition of substantially varying languages
to related dialects exhibiting only superficial differences. Rather surprisingly,
Greek theorizing and the Greek dialect context did not play a fundamental role
in the emergence of this phrase, even though it might have been indirectly
inspired by early sixteenth-century discourse on Greek grammar and by the
traditional definition proprietas linguae, a Latin calque of Greek idíōma glōś sēs.
The emergence of the collocation was principally urged by a need for mapping out
the close relationships between different peoples, especially those that were lesser-
known, such as the inhabitants of ancient Gaul. Tellingly, I have not encountered
any instances of authors claiming that the Greek tribes differed only in dialect,
most likely because such a claim would have been perceived as stating the obvious.
5.6 An intricate conceptual web
The sixteenth-century establishment of the language/dialect opposition was not

a straightforward process, since dialect interfered and interacted with other
concepts such as idiom and style in an utterly complex and elusive way. In
particular, the intricate relationship between dialect and idiom, in the sense of
‘the specific character or individuality of a language, the manner of expression
considered natural to or distinctive of a language, a language’s distinctive
phraseology’, surfaced in late fifteenth- and sixteenth-century definitions of
dialectus (diálektos) and idioma, both terms with several different meanings.²⁴
The former word was frequently explained as idioma loquendi, ‘property of
speaking’, while many authors defined the latter as proprietas sermonis, ‘property
²⁴ The definition of idiom is from the online Oxford English Dictionary, s.v. idiom (last accessed
8 April 2019).
.     77
of speech’.²⁵ The definition proprietas linguae, ‘property of tongue’, was even

used to explain both dialectus and idioma.²⁶ The terms were, in other words,
often perceived as synonyms, or at least near-synonyms. One could, in fact, say
that dialectus was an alternative terminological means to express the concept of
idiom, which was not as clearly distinguished from dialect as it is in present-day
linguistics. As I have argued in Chapter 3, Roger Bacon had already confused the
two concepts in a similar manner in the thirteenth century.
In the course of the sixteenth century, the word dialectus was singled out as the
term most adequate to render the concept of dialect, ‘variety of a language’. The
reasons behind this development are not entirely clear. This selection process
moreover appears to have occurred undeliberately, as I will argue in the next
chapter. It might be related to the fact that the term diálektos, unlike idíōma, had
been explicitly and repeatedly associated with particular properties such as ‘spe-
cific to a certain tribe’ and—especially—‘specific to a certain place’ in Greek
scholarship. This connection would have made dialectus and, later on, its ver-
nacular equivalents more apt candidates to express the concept of dialect than
idioma and its vernacular forms, which lacked a close association with such
specific parameters. I would, however, not go so far as to claim that idioma was
never used to convey the concept of dialect; it was just not the preferred option.
The ongoing semantic specialization of the term dialectus did not, however,
prevent a confusing usage of it. A remarkable case in point is Conrad Gessner’s
teacher Theodore Bibliander, who operated with a quite blurred conceptual
apparatus. In several cases, Bibliander seems to have been opposing dialectus to
lingua. He stated, for instance, that ‘all dialects of the Greeks are called one Greek
language’.²⁷ But elsewhere he confused the concepts of dialect and idiom, most
conspicuously when he claimed:
I see that some ancient grammarians find it pleasing to call the “idiom” of a
language the form of speaking which emerges from the structure of words. For
instance, what the Greeks say as “hépomai soi” [‘I follow you’: verb + dative], the
Latins say as “sequor te” [‘I follow you’: verb + accusative]. And “mákhomai
mákhēn”, ‘I fight a fight’, is called Attic diction. But the matter manifests itself
more widely, which is both noticed by reasoning and established by authority.
For John the Grammarian, who has written on the dialects of the Greeks, places
the difference of tongues either in the entire word—for instance, what the Ionians
say as “kárta” [‘very; exceedingly’], the other Greeks say as “lían”—or in a part of
the word—the Aeolians, for instance, say “bróda” [‘rose’], the other Greeks
“rhóda”—or in the affection, or passions, of a word—the Dorians, for instance,
²⁵ See e.g. Crastone ([1478]: s.v. διάλεκτος) and Calepino (1502: s.v. idioma), respectively.
²⁶ See e.g. Sabellicus (1490: 64) and Estienne (1531: s.v. idioma).
²⁷ Bibliander (1542: 58): ‘omnes Græcorum dialecti una lingua gręca dicuntur’.
78   
pronounce “pantôs” [‘at any rate’] with a circumflex accent, which the others
pronounce with a grave accent as “pántōs”.²⁸
Likewise problematic for sixteenth-century scholars was the relationship

between dialect and style, in the sense of ‘the manner of expression characteristic
of a particular writer (hence of an orator), or of a literary group or period’,
especially since philological interests came to influence interpretations of the
dialect concept.²⁹ I have indicated earlier how Quintilian compared Latin literary
registers to the Greek dialects (Chapter 2, Section 2.4). In the sixteenth century,
the confusion of both concepts can be perceived in, for instance, the humanist
Joachim Camerarius’s (1500–74) proem to his edition of Herodotus’s Histories of
1541. In his paragraph On the Ionic dialect, Camerarius stated that ‘the dialect, i.e.
peculiar kind of discourse, of the Ionians has a great conformity with the Attic
tongue’, after which he related where and by whom the Ionic dialect was spoken.³⁰
Both his paraphrase of dialectus as ‘peculiar kind of discourse’ and his character-
ization of Herodotus’s style as ‘elegant’ and ‘exhibiting a certain fluency particular
to Ionic’ seem to betray a certain confusion of dialect with style.³¹
The ambiguous place of dialect in this intricate conceptual web had implica-
tions for its interpretations. As a matter of fact, the polysemous term dialectus was
often used in such an unstraightforward manner that its meaning needed to be
guessed from its immediate context, thus leaving ample room for each reader to
interpret the word as it suited them. This interpretive freedom seems to have
greatly contributed to, among other things, the idea that a dialect derived from,
and was posterior to, a language, which became popular mainly after 1600 (see
also Chapter 10). Such a diachronic conception of the conceptual pair was,
however, already implicit in the usage of Conrad Gessner, who represented ‘the
dialects of the Germans’ as ‘descendants of the same original language’.³² Whence
did Gessner draw inspiration for this assumption? I argue that his reading of his
teacher Theodore Bibliander’s work is likely to have stimulated his at times
genealogical usage of the term dialectus, even though this interpretation did not
necessarily correspond to Bibliander’s intentions. When discussing linguistic
²⁸ Bibliander (1542: 89): ‘Ιδίωμα linguæ dici loque[n]di formulam, quæ ex uerboru[m] structura
nascitur, uideo placere ueteribus quibusdam gra[m]maticis, ut quod Græci dicu[n]t ἕπομαι σοι, Latini
sequor te. Et Attica phrasis dicitur μάχομαι μάχην pugno pugnam. Sed latius rem patere tum ratione
animaduertit[ur], tum authoritate comprobatur. Nam Ioannes Gra[m]maticus, qui de Græcorum
dialectis scripsit, uel in tota collocat dictione linguaru[m] differentiam, ut quòd Iones dicunt κάρτα,
reliqui Gręci λίαν, uel in dictionis parte, ut Aeoles βρόδα dicunt, reliqui Græcorum ῥόδα: uel affectione
uocis, aut passionibus, ut Dores παντῶς circunflexo tono efferunt, quod alij graui tono πάντως’.
²⁹ See the online Oxford English Dictionary, s.v. style (last accessed 8 April 2019).
³⁰ Camerarius (1541: β.2): ‘Dialectus autem, id est, peculiare orationis genus, Ionum cum Attica
lingua magnam habet communitatem’.
³¹ For the common transfer of textual properties to language varieties see Schlieben-Lange (1992).
³² Gessner (1561: *.5): ‘sanè Germanorum dialecti per uniuersum Rhenum à fontibus eius ad
Oceanum usque, omnes eiusde[m] linguæ principalis ceu propagines sunt’.
.  79
monogenesis, for instance, Gessner’s teacher stated that God acted ‘so that out of
one common speech several dialects emerged’.³³ Elsewhere, Bibliander offered a
more extensive description of his monogenetic approach to language history:
Since Hebrew is the primeval language, the others are generated and born out of
it. Some of these degenerated farther from the original speech, such as the
dialects of Japheth’s sons. Others kept closer to the first language, such as
the dialects of Shem’s sons. Still others occupied a position in-between, such
as the dialects of Ham’s sons.³⁴
Confronted with such passages, readers such as Gessner might have intuitively
understood dialectus as a ‘variety derived from a chronologically primary lan-
guage’, even though such an interpretation is simply impossible in other passages
of Bibliander’s work, for example, where he observed that ‘the books of all Jewish
writers are filled with various dialects: Hebrew, Syriac, Arabic, German, Greek,
Latin, Italian, French, Spanish, and what not?’³⁵ Its meaning here remains unclear,
but dialectus is surely not to be taken in a genealogical sense. Bibliander’s (1548:
107) discussion of the term dialectus does not provide any clues in this regard, as it
is rather a state of the art of Greek usage of the term than an explanation of his
own conceptions. Several meanings were listed—‘property of tongue’, ‘language’,
‘speech’, ‘rare word’, etc.—but he did not express his preference for any of them.
Readers are left to themselves when it comes to determining what Bibliander
precisely meant by dialectus in each context and are free to interpret the term as it
suits them.
5.7 Conclusion
In the years 1500–50, a distinction between language and dialect emerged, intro-
duced primarily by scholars active in German-speaking areas, with a key role for
scholars of Swiss origin such as Theodore Bibliander, Conrad Gessner, and
Johannes Rhellicanus. They were motivated to do so by a fascination with the
phenomenon of human language and its diversity, often connected with an
interest in ancient history.
³³ Bibliander (1542: 52): ‘ut ex uno sermone co[m]muni, plures dialecti prodire[n]t’.
³⁴ Bibliander (1548: 142): ‘Siquidem Ebræa est primigenia, reliquæ ex ea propagatæ & genitæ sunt.
Quarum aliæ longius degenerarunt à principali sermone, ut dialecti filiorum Iapheth: quæda[m] arctius
adhæserunt linguæ primæ, ut dialecti filiorum Sem: quædam medio se modo habent, ut dialecti
filiorum Cham’.
³⁵ Bibliander (1542: 82): ‘Omniumq́[ue] Iudaicoru[m] scriptoru[m] libri uarijs dialectis referti sunt,
Hebręa, Syra, Arabica, Germanica, Gręca, Latina Italica, Gallica, Hispanica, & qua no[n]?’
80   
Even if the emergence of the conceptual opposition was not reflected in a

consistent metalanguage, it cannot be denied that lingua and dialectus were
increasingly employed contrastively from the early sixteenth century onwards.
The terminological distinction manifested itself in different ways. Apart from a
direct and explicit opposition of the two words, sixteenth-century scholars also
developed the concept of common language, lingua communis in Latin. This
concept presupposed the existence of non-common varieties, and we indeed
find that lingua communis was opposed to dialectus by prominent scholars such
as Philipp Melanchthon. Other main manifestations of the lingua/dialectus con-
trast are the addition of determiners to lingua in definitions of dialectus and the
coining of phrases such as ‘dialect(s) of language x’ and ‘to differ only in dialect’,
which was usually taken to imply ‘and not on the level of language’.
The contrastive usage of lingua and dialectus moreover reveals that several
criteria came to be associated with the two concepts, albeit only loosely at first.
Various scholars suggested, for instance, that related dialects exhibited superficial
differences, whereas distinct languages were marked by substantial variation. It is,
however, not easy to define a core dialect concept for the early sixteenth century.
If, however, I were to formulate one, I could circumscribe dialect as ‘particular
speech form subsumed under a certain language’. This definition would have the
advantage of emphasizing a sixteenth-century development with major repercus-
sions for the later history of linguistic thought and praxis: the conceptual sub-
sumption of one linguistic entity, a dialect, under another, a language. Such a
formulation nevertheless remains undesirable in my opinion, as it does great
injustice to the versatility with which sixteenth-century scholars relied on the
dialect concept to attain their intellectual and argumentative goals. The definition
could moreover convey the misleading impression that dialect was a clearly
distinct and uniform concept, but the opposite is true. Its elusive interference
with other concepts as well as the polysemy of the term dialectus greatly hampered
more concrete and better-defined interpretations of the concept in the first half of
the sixteenth century. Scholars were, in other words, unable to develop a man-
ageable concept allowing them to account for the minor differences they noticed
within languages that were perceived as otherwise uniform.
It is, finally, remarkable that the emergence of the conceptual pair followed the
borrowing of Greek diálektos into Latin fairly quickly, which suggests that the
borrowing of dialectus can be taken to forebode intellectual innovation and that
scholars were looking for the necessary metalanguage to talk about language-
internal variation in the scientific lingua franca of the Renaissance. Lexical bor-
rowing can in this case be viewed as paving the way for conceptual change. One
key issue relating to the sixteenth-century genesis of the language/dialect distinc-
tion remains to be addressed: why did it occur as it did, and in which context?
6
Hellenism, standardization, and info-lust
The genesis of the conceptual pair in context
A considerable number of sixteenth-century scholars suddenly began to assume

that the Latin terms lingua and dialectus referred to linguistic entities of different
orders. Which contextual factors contributed to the increasing contrastive usage
of these two terms? And were humanist scholars consciously innovating, as Roger
Bacon was in the thirteenth century?
6.1 A major pivoting point: the rediscovery

of the Greek dialects
The countless references to the Greek heritage in the two preceding chapters are
no coincidence; the rediscovery of the Greek language and literature and, espe-
cially, of their formal plurality was a major factor in the emergence of the
language/dialect distinction.¹ As a matter of fact, contrasting lingua and dialectus
became common practice in the first place among scholars trained as Hellenists,
specialists of the Ancient Greek language and literature. The term dialectus was by
virtue of its origin closely associated with the Greek language and its different
forms, at least initially. Indeed, in most available Greek texts, diálektos referred to
a variety of the Greek language and also in the rare instances in which it occurred
in Latin texts, it almost exclusively referred to forms of Greek. This semantic focus
explains why dialectus was initially only part of the vocabularies of Hellenists,
nearly always with reference to the Greek language. Where did the sixteenth-
century interest in the Greek dialects come from? And why did it emerge only
then and not earlier, say, after the Byzantine diplomat Manuel Chrysoloras
(c. 1355–1415) started teaching Greek in Florence from 1397 onwards?
Despite the obvious dialectal diversity present in the literary Greek texts that
were being rediscovered on the Italian peninsula, and despite the importance of
Greek for the emancipation of the dialectally diverse Italian vernacular,
Quattrocento Hellenists showed remarkably little interest in the phenomenon of
¹ A more detailed account of the early modern rediscovery of the Greek dialects is in Van Rooy
(2020a), on which I draw here.
DOI: 10.1093/oso/9780198845713.001.0001
82 , ,  -
dialectal variation.² Why was this the case? When Byzantine émigrés started
teaching Greek in Italian city-states, they had to focus on the basics of grammar
and were not in a position to trouble their students too much with the complicated
matter of the dialects. That is not the entire explanation, however. The Byzantine
teachers do not seem to have been experts in the issue either. In 1441, the Italian
humanist Francesco Filelfo (1398–1481) lamented that even in Constantinople,
where he had lived for several years between 1420 and 1427, no Aeolic was taught
(Rotolo 1973–4: 88 n.4; cf. Botley 2010: 71–114). What is more, Greek treatises on
the dialects seem to have been largely unknown before the end of the fifteenth
century. As a result, texts in certain dialects, mainly Aeolic and to a lesser extent
Doric, could not be easily studied during the Quattrocento.
Does that mean that Greek dialectal features were not treated at all in the
didactic texts composed in the fifteenth century by Byzantine migrants for their
Western students?³ Some of them did indeed insert some remarks on the par-
ticularities of Attic, Ionic, Doric, and Aeolic. Chrysoloras, for instance, probably
made limited use of Gregory of Corinth’s treatise on the Greek dialects when
composing his grammatical handbook Erōtē m ́ ata (Ἐρωτήματα), Questions
(Botley 2010: 166 n.70). In fact, he mentioned idiosyncrasies of the canonical
dialects throughout this work.⁴ Something similar occurred in grammars of later
Byzantine refugees, for instance, in Theodore Gaza’s (c. 1398–c. 1475) four books
on Greek grammar, composed around 1461–2.⁵ It seems that Byzantine scholars,
as a general rule, eclectically referred to what they perceived to be the most notable
dialectal properties. An exception to this tendency was Constantine Lascaris
(1434–1501), who authored a treatise in which dialectal and poetical variation
in the Greek pronoun was discussed more systematically.⁶
It was only in the grammatical work of Western Hellenists that more attention
was devoted to the phenomenon of the Greek dialects. The different forms of the
language were especially prominent in grammars from the late fifteenth century
onwards. The Italian humanist Urbano Bolzanio (1442–1524) is a telling case, as
he clearly struggled with the subject of the dialects and its presentation throughout
the different editions of his handbook (Botley 2010: 36–40). The dialects occupied
an ever more prominent position in publications on the Greek language during
the first decades of the sixteenth century, at the exact same time when the
conceptual pair emerged. Tellingly, this period witnessed the printing of separate
handbooks for the Greek dialects. The first in its kind was a booklet entitled On the
² On the importance of Greek for the emancipation of the Italian vernacular see e.g. Dionisotti
(1968: 51).
³ On these handbooks in general see Botley (2010), with further references.
⁴ See e.g. Chrysoloras (1512: 20). I refer to an early modern edition, as a modern critical edition is
still lacking.
⁵ See e.g. Gaza (1495: г.v) for a Doric particularity. Pace Botley (2010: 17).
⁶ Lascaris (1502: z.viii). See Botley (2010: 26, 124, 175 n.272).
.     83
Figure 6.1 The first separate edition of Adrien Amerot’s On the diverse dialects, 1530
Source: KU Leuven Libraries. Public domain
diverse dialects of Greek declinations, both in verbs and nouns (Figure 6.1; Amerot
1530). The text was printed in 1530 in Paris, at the atelier of Gérard Morrhy, and
was an excerpt from a Greek grammar published in Leuven ten years earlier: the
Compendium of Greek grammar by Adrien Amerot (c. 1495–1560) from Soissons,
later professor of Greek at the Collegium Trilingue in Leuven (Amerot 1520).
84 , ,  -
Not only did Amerot’s booklet become a bestseller, enjoying countless reprints
during the entire early modern period (Hoven 1985: 5–19); it also initiated a
tradition of printed manuals for the Greek dialects lasting up to the present day.
In sum, the Greek dialects took up an increasingly visible space in the intellec-
tual landscape of the early sixteenth century. Hellenists devoted ever more atten-
tion to the subject and started applying the terms they found in Greek
grammatical works to other linguistic contexts as well, most often their native
vernacular one. This terminological extrapolation first and foremost involved
Latin dialectus, but also the Latin rendering of Greek koinē ́ as communis lingua,
as I have contended in the previous chapter. The learned status of the Greek
language was of central significance in this regard. Its variation could impossibly
be neglected by grammarians and philologists and often led them to reflect,
however rudimentarily, on their native dialect context, a development that
would likely have been delayed without the renewed interest in the Greek heritage.
In other words, the multidialectal outlook of sixteenth-century Hellenists trig-
gered intuitive comparisons of different dialect contexts, usually the Greek and
their native ones, leading them to develop an abstract conceptual distinction
between language and dialect.
The rehabilitation of the Greek heritage is not the entire explanation, however,
and there are additional circumstances that considerably contributed to the
emergence of the conceptual pair. What were these? I argue that the standardiza-
tion of vernacular tongues and the information explosion with which humanist
scholars were confronted were of profound significance as well.
6.2 From ancient Greece to Western Europe:

standardizing the vernaculars
In the Renaissance, the vernacular tongues of Western Europe were on the rise,
partly as a result of the changing social order. The merchant middle class in
particular prospered, who felt barely any affinity with Latin, the scholarly lingua
franca, and used their native tongues to arrange their businesses. The promotion
of the vernaculars at the same time entailed strong competition between them.
Humanists such as Joachim du Bellay made a point of defending and elevating
their native language, French in du Bellay’s case (cf. Burke 2004: 61–88). This
linguistic competition should be framed within the humanists’ desire to set
their own imagined national community apart from others.⁷ In doing so, human-
ists tended to put an emphasis on the borders between peoples and countries as
⁷ On the intensifying link between language and national communities in the early modern era see
Joseph (2006: 22).
.       85
well as on their own superiority in political, cultural, and moral terms. This kind
of profiling was a reflection of their political engagement and aspirations. Put
differently, emerging national sentiments and pride led humanists to accord
higher status to their native vernaculars, contending that they excelled their
neighbours also in linguistic terms.
The renowned political scientist and historian Benedict Anderson (2006:
37–46) has pointed out in his well-known Imagined Communities the importance
of print-capitalism for the emergence of national consciousness to the detriment
of ‘the imagined community of Christendom’. Anderson connected this new
consciousness not only with the development of humanist Latin and the rise of
Protestantism, but also, and most notably, with the promotion and standardiza-
tion of vernacular tongues. However, this national turn occurred less self-
consciously than in nineteenth-century nation-states, Anderson argued, and
linguistic standards were in the Renaissance never imposed on a state’s entire
population. Along similar lines, Caspar Hirschi (2012: 13, 41, 109, 126) has traced
the origins of nationalism back to humanist discourse and sees an important role
for linguistic factors. According to John E. Joseph (2004: 48, 107), ‘national
language’ can even be regarded as a ‘dream of the sixteenth century’, with
linguistic standardization being primarily motivated by a desire to emancipate
the vernacular from Latin. In view of the emerging national consciousness in
sixteenth-century Western Europe, it is no coincidence that this period witnessed
the first attempts at systematically associating national groups with stereotypical
characteristics (Leerssen 2007: 17).
What were the implications of the competition among states and their national
vernaculars? This linguistic race presupposed a clear demarcation of one’s own
language from other vernaculars and went hand in hand with standardization
processes. John Considine (2008a: 104) has aptly summarized it: ‘once one variety
became recognized as a standard, it could become a symbol of membership of a
wider community and of difference from other communities’. The competing
vernacular tongues were identified with the selected varieties which were being
codified and elaborated and to which all other dialects were subordinated (Joseph
2006: 29–30). The process of linguistic selection and especially the hierarchical
subordination inherent to it no doubt catalysed the emergence of the conceptual
pair, all the more since ‘standards, by inviting comparison of linguistic forms,
enable conceptualization of non-standard varieties’ (Machan 2003: 98).
The impact of two historical circumstances deserves to be emphasized when it
comes to the rise of the vernaculars in relation to the genesis of the conceptual
pair: the printing press (Figure 6.2) and the Reformation. The appearance of the
printing press greatly contributed to the codification of vernacular standard
languages, as it encouraged fixing their form (cf. Anderson 2006: 45). The printed
word enhanced the formal stability of the vernaculars and intensified the contrast
between uniform printed language and the great variability of spoken dialects.
86 , ,  -
Figure 6.2 Woodcut showing sixteenth-century printing in action, 1568

The Reformation, in turn, led to the promotion of the vernacular, especially in

Germany, where Luther became a bestselling scholar with, among other things, his
German Bible translation (Padley 1988: 246). Coincidentally, Protestant scholars
developed a greater interest in dialectal variation than their Catholic colleagues
(cf. Van Rooy 2020c).
.   87
6.3 Knowledge revolution: information explosion

and info-lust
Sixteenth-century scholars were moreover confronted with an ‘information

explosion’ and concomitantly driven by a great ‘info-lust’, to use two phrases of
Ann Blair (2010: 11–12). In other words, scholars faced large amounts of data and
felt the desire to gather and organize them, for which they developed a wide-
ranging apparatus of categories. The Renaissance collector’s mania was probably
embodied best in the emergence of curiosity cabinets, rooms where all kinds of
rare objects were gathered and which prefigured later museums (Figure 6.3; see
e.g. the studies in Impey and MacGregor 2001). Languages, impossible to capture
live in their spoken form during the Renaissance, were collected through the
medium of written and printed texts in libraries.
Assiduous attention to empirical data went hand in hand with a critical
approach to ancient and medieval texts, with more and more humanists ques-
tioning the authority of classical authors by looking at what was right in front of
their noses. The emergence of the conceptual pair can be seen as a product of the
humanist info-lust, or rather as a side effect of it, since it was considered to be a
self-evident given and not a newly conceived metalinguistic tool, as I will argue in
Figure 6.3 Natural history cabinet, 1599

88 , ,  -
Section 6.5. The distinction between language and dialect helped scholars to map
out and make order in the enormous linguistic diversity that became ever more
obvious in the sixteenth century.
It is certainly no coincidence that the first extensive language catalogue ever
composed, Conrad Gessner’s Mithridates of 1555, was published in the same
period in which the language/dialect pair emerged and became common as a
metalinguistic distinction. Gessner might even have conceptualized his book as
a kind of linguistic curiosity cabinet, in which he put the conceptual pair into
practice as a welcome means to group and categorize the phenomenon of lan-
guage diversity (cf. Sergeev 2019).
6.4 The countability of language
The historical circumstances sketched above are linked to a general feature of

sixteenth-century linguistic thought: the tendency to conceptualize languages as
clearly distinct, enumerable entities, as pointed out at length by the Dutch linguist
Joop van der Horst (2008: 136–78). This conception surfaces, first, from the idea
that different languages and the variation in each of them could be compared.
Second, the ongoing standardization of, and competition among, vernacular
languages involved the design of a clearly distinct status and profile for each of
them. Thirdly, humanists worked on the basic assumption that languages were
objects of study that could be categorized.
While it is true that, before the early modern period, many scholars supposed
on the authority of the Bible that there was a fixed number of languages, usually
seventy-two, few of them were able to actually enumerate them all. Furthermore,
counting languages was not an easy endeavour, owing to the fact that many
languages consisted in continua of varieties without a recognized standard, let
alone clear boundaries. This state of affairs changed in the sixteenth century and
had a profound impact on the metalinguistic conceptual apparatus. Indeed, the
emphasis on the idea that languages were countable and demarcated seems to
have been an important trigger for scholars to develop the language/dialect
distinction. While establishing ever clearer boundaries between languages, fre-
quently associated with political borders, it became ever easier and ever more
obvious to record variation within one language.
The example of the Greek language and its five principal dialects stimulated the
idea that vernacular dialects, too, were distinct entities and were therefore count-
able. As a matter of fact, several Hellenists started enumerating vernacular dialects
from the early sixteenth century onwards, often assuming the same number for
the Greek and the vernacular dialects. The French scholar Geoffroy Tory (1480–c.
1533) provided an early example, when he noted in his Flowery field of 1529, a
manifesto for developing a French standard language:
.     89
Our language is as easy to regulate and put in good order as the Greek tongue
once was, in which there are five diversities of language, which are the Attic, the
Doric, the Aeolic, the Ionic, and the Common tongue, which have certain
differences among each other in the declensions of nouns, in the conjugations
of verbs, in orthography, in accents, and in pronunciation, as a Greek author
called John the Grammarian and several others treat and teach very extensively.
Exactly that we would well be able to do for the tongue of the court and of Paris,
the Picard tongue, the Lyonnais, the Limousine, and the Provençal. I would
discuss some of their differences and correspondences, except that I do not want
to be too long here and I leave it to those who are more experienced than me to
devote themselves to it.⁸
Some humanists even pointed out that Greek had fewer dialects than their
native tongue (e.g. Wolf 1578: 595 on German dialects). What is more, from the
early seventeenth century onwards, scholars designed classifications of the prin-
cipal dialects of a vernacular language, often inspired by the Greek example. The
English schoolmaster and grammarian Alexander Gill (1565–1635), for instance,
described his native context as follows in Chapter six of his English grammar,
entitled ‘Dialects, in which the improper diphthongs are also treated’:
There are six principal dialects: Common, Northerners’, Southerners’,

Easterners’, Westerners’, Poetic. I neither know nor have heard all their particu-
larities. Yet I will describe them as far as I can those I remember.⁹
Here, Gill enumerated the most important English dialects, and several elements
suggest influence from early modern grammars of Greek, with which Gill must
have been acquainted. Not only was he a distinguished Hellenist—he taught the
language to John Milton—but he also presupposed the existence of a common
English variety, dubbed dialectus communis, a designation no doubt inspired by
the Greek phrase koinē ́ [diálektos] (Van Rooy 2020a: 23).
The assumed countability of dialects went hand in hand with a terminological
development, as dialects were increasingly accorded distinct labels based on
⁸ Tory (1529: –): ‘Nostre langue est aussi facile a reigler et mettre en bon ordre, que fut iadis la
langue Grecque, en la quelle ya cinq diuersites de la[n]gage, qui sont la langue Attique, la Dorique, la
Aeolique, la Ionique, & la Comune, qui ont certaines differences entre elles en Declinaisons de noms, en
Coniugatio[n]s de verbes, en Orthographe, en Accentz & en Pronunciation. Co[m]mme [sic] vng
Aurheur [sic] Grec nomme Ioa[n]nes Gra[m]maticus, & plusienrs [sic] autres traictent & enseignent
tresamplement. Tout ainsi pourrions nous bien faire, de la langue de Court & Parrhisiene, de la
la[n]gue Picarde, de la Lionnoise, de la Lymosine, & de la Prouuensalle. Ien dirois aucunes differences
& accordances/se nestoit que ie ne veulx icy estre trop long, et que ie laisse a plus expertz que moy eulx
y employer’.
⁹ Gill (1619: 15): ‘Dialecti: vbi etiam de diphthongis improprijs. Dalecti præcipuæ sunt sex:
Communis, Borealium, Australium, Orientalium, Occidentalium, Poetica. Omnia earum idiomata
nec novi, nec audiui; quæ tame[n] memini, vt potero dicam’. See Kökeritz (1938).
90 , ,  -
toponyms, usually adjectives referring to a city or region, as the examples of Tory

and Gill indicate. This heightened labelling activity might likewise have been
partly inspired by the Greek tradition, especially since the Greek dialects were
exceptional in that they received a name early on in antiquity (Adams 2007: 11).
6.5 A product of appropriation and subconscious adaptation
I have been calling the emerging terminological contrast between lingua and
dialectus a ‘symptom’ of the genesis of the conceptual pair language/dialect,
which gives a clear hint about the way in which the distinction came into being.
As a matter of fact, sixteenth-century humanists introducing the lingua/dialectus
distinction were largely unaware of their scholarly achievement, since they seem to
have assumed that they were simply using the term dialectus just as the ancient
Greeks had used diálektos. They were, however, reading their own interpretation
into the term, a process involving an opposition of dialect to language. The
conceptual pair, in other words, came about in a subconscious way, almost by
accident, as it were. This uncoordinated emergence stands in startling contrast to
the case of the late medieval polyglot Roger Bacon, who was clearly aware that his
lingua/idioma pair was an innovation. As I have contended, Bacon conceived it as
a reaction against the all too liberal usage of the term lingua by his colleagues.
There is a striking parallel with the history of the word dialect here. Just like the
emergent opposition of lingua to dialectus, the Latinization of the Greek word
diálektos was perceived as a realization of ancient, this time Roman, scholarship,
even though fifteenth- and sixteenth-century humanists were actually reading the
word dialectus into Quintilian’s work, but mistakenly so. Briefly, the appropri-
ation of Greek diálektos went hand in hand with subconscious conceptual adap-
tation by sixteenth-century scholars. Inspired by Ashley and Plesch (2002),
I understand appropriation here as a complex process of adopting and adapting
an existing concept, either consciously or subconsciously, to one’s particular
historical context and discourse. The subconscious nature of the appropriation
of the Greek word makes it impossible to name the prôtos heuretē ś , the ‘first
inventor’, of the distinction between lingua and dialectus. This historical fuzziness
reminds in some ways of the mid-fifteenth-century invention of the printing
press.¹⁰ Let me, by way of thought experiment, consider three major points of
comparison.
First, just as printing with movable type was perfected by the combined efforts
of Mainz printers, the conceptual pair did not emerge as the result of a one-man
endeavour, but was the product of a specific social group: the early sixteenth-century
¹⁰ The information on the invention of the printing press is based on Rice and Grafton (1994: 1–10).
.  91
Republic of Letters, and, as I have argued, especially those members that were
skilled in the Greek language. It is, however, not impossible to name a number of
key figures. For printing, those were Johann Gutenberg (c. 1395–1468), Johann
Fust (c. 1400–1465), and Peter Schöffer (c. 1425–1502). For the language/dialect
distinction, one could mention, among several others, Philipp Melanchthon,
Juan Luis Vives, Petrus Antesignanus, and Conrad Gessner. The humanists
pioneering the conceptual pair were, however, not as geographically concen-
trated as were the Mainz printers, but this is related to the international
character of the Republic of Letters and the mobility of its members.
Second, both the printing press and the language/dialect distinction display the
typical Renaissance blend of old and new. On the one hand, the invention of
printing was much indebted to the import of two Chinese ingenuities, block
printing and paper, and to existing European techniques such as oil-based ink,
which were combined and used in innovative ways. The conceptual pair, on the
other hand, was the result of the humanist appropriation of Greek scholarship on
the literary dialects and the concomitant adaptation of Greek definitions of
diálektos. They subconsciously tailored the term to the linguistic realities of
sixteenth-century Western Europe and their outlook on them.
Thirdly, Gutenberg and his colleagues ‘considered printing only a new and
particular kind of writing’, which reveals ‘[t]heir difficulty in freeing themselves
from traditional conceptions’ and indicates that they were not aware of ‘the
unique potentialities of their invention’, to put it with the words of Rice and
Grafton (1994: 5). The printers’ lack of perspective has an intriguing parallel in the
humanists’ unawareness of the fact that, instead of simply furthering Greek
tradition, they were actually making a contribution to linguistic thought that
was to exert an impact on scholarship lasting up to this day.
It is, however, not only points of comparison which link the printing press and
the language/dialect distinction. The production and distribution of printed books
‘turned intellectual work as a whole into a cooperative instead of a solitary human
activity’, to quote Rice and Grafton (1994: 8) once more. This very fact no doubt
accelerated the emergence of the conceptual pair, which cannot be attributed to
one single humanist.
6.6 Conclusion
The conceptual pair language/dialect emerged as the result of a complex interplay

between several aspects of early modern scholarship and society: the rediscovery
of the classical Greek heritage, the standardization of vernacular tongues, and the
overwhelming amounts of new information with which scholars were confronted.
The language/dialect distinction was especially prominent in the works of Italian-
and German-speaking scholars, who were the first to appropriate Greek studies
92 , ,  -
and for whom standardization was an acute problem due to the absence of a firmly
established centralized state and, in the case of Germany, due to the Protestant
insistence on adequately translating the Bible into the vernacular. The emergence
of the conceptual pair was therefore most likely the result of an adjustment to new
linguistic realities as well as of a growing interest in linguistic diversity on the part
of sixteenth-century scholars, who made ambiguous use of the Greek heritage in
doing so.
It may also be stressed that the emergence of the conceptual pair was far from
a straightforward process. Its most striking feature is that it was created not
rectilinearly but almost by accident as it were, with scholars assuming they were
just continuing Greek tradition. They simply took the existence of a language/
dialect distinction for granted, failing to recognize its absence from earlier
thought. Additionally, conceptions of dialect as subsumed under language
remained rather vague in the first half of the sixteenth century. As a result, it is
unclear whether sixteenth-century authors conceived of a dialect as a specific
subsystem of a language or as a loose collection of regional features within a
language. I am inclining towards the latter answer, since dialectal variation was
usually approached in a highly punctual manner, that is, principally in terms of a
range of inconsistent letter changes. Whatever the case, it is clear that defining
what a dialect exactly was seems to have been a frustrating endeavour for many
sixteenth-century scholars. It was only after about 1550 that more elaborate
interpretations began to appear, as I set out to argue in Part III.
III
C O N S O L I D A T I O N BY
E L A B O R A T I O N , 1 5 5 0 – 1650
7
Space and nation
Greek definitions transformed
In the first half of the sixteenth century, Hellenist humanists developed a language/
dialect distinction. They did so independently from Roger Bacon and subcon-
sciously, as they assumed they were just following in the footsteps of their great
Greek examples. As a result, they did not feel the need to make explicit what they
understood by the conceptual contrast, at least at the outset. After about 1550,
their attitude began to change, and scholars increasingly formulated explicit
interpretations of the distinction and especially of the dialect pole, thus consoli-
dating their innovation by elaboration. What interpretations did they propose?
It is this question which I address in the current and following chapters. Even
though the focus is on the years 1550–1650, I make occasional digressions into
other time frames to highlight certain developments that are of interest to the
interpretation in question. I start this chapter with two conceptions with firm
roots in Greek thought: dialect defined in terms of space and nation.
7.1 The spatial conception of dialect
[A dialect is] a particularity of some one and the same language,

which is different for different provinces or cities using that language.¹
This is how the Wittenberg professor of Greek, Hebrew, and mathematics

Erasmus Schmidt (1570–1637) defined his usage of the term dialectus in his hand-
book on the Greek dialects, published in 1604. Apparently, Schmidt conceived of a
dialect as a variety of a language that was in restricted use in regional terms, which
implied that a language covered a larger area. When did this geographical interpret-
ation of the conceptual pair emerge? It seems to have appeared only in definitions of
dialectus formulated after 1550. An early mentioning of geographical restriction in a
definition of dialectus in contrast with lingua occurred in Petrus Antesignanus’s
widely read and often reprinted commentary on the Greek grammar by Clenardus.
In this work of 1554, Antesignanus defined dialectus as
¹ Schmidt (1604: 4): ‘idioma unius & eiusdem alicuius linguæ quod diversis Provinciis vel
Civitatibus eâ utentibus, est diversum’.
DOI: 10.1093/oso/9780198845713.001.0001
96   
a certain property of speaking, by which the variety of speech is distinguished.

This usually occurs between different regions of the same nation, because some
speak a little differently from others and something proper from the native soil is
engrafted in everyone.²
In contrast, Antesignanus assumed a common language to supersede this restrict-

edness in regional terms (cf. Chapter 5, Section 5.3).
The absence of the geographical interpretation from earlier understandings of
the conceptual pair probably relates to the fact that Clement of Alexandria’s
Stromata were only published in 1550 in Florence. In this text, the Greek word
diálektos was broadly defined as speech that bore the mark of a certain locality,
tópos in Greek. Clement, however, had not yet subsumed diálektos under a
language-like entity. Antesignanus did not mention Clement’s definition, but it
is very likely no coincidence that Antesignanus’s geographical interpretation of
the conceptual pair roughly coincided with the first printing of Clement’s work.
The second author in whose work the spatial conception of dialect appeared did
extensively refer to and even quote Clement’s Stromata, and this only one year after
Antesignanus’s work was published. I am, of course, referring to Conrad Gessner,
whose reliance on Clement I have analysed at greater length in Chapter 4. Not only
the availability of Clement’s work in a printed form since 1550, but also especially
the success of Antesignanus’s grammatical commentary of 1554 and Conrad
Gessner’s Mithridates of 1555 allowed for the rapid diffusion of the geographical
interpretation of dialectus as opposed to lingua. Both works were widely read and
enjoyed reprints, especially Antesignanus’s work, part of a popular manual for the
Greek language. In the sixteenth century alone, there were at least fifty-five editions
of the book across Western Europe (Van Rooy 2016c: 134).
It cannot, however, be denied that variation in a language was associated with
locality even before the 1550s, albeit on a lower level of abstraction. Notably, the
French philosopher Charles de Bovelles (c. 1479–after 1566) drew up an impres-
sive account of Gallo-Romance regional diversity in his Book on the difference of
vernacular tongues, where he tied the phenomenon to the prevalence of numerous
local legal systems (coutumes; Bovelles 1533), a crucial step in recognizing lin-
guistic difference, standardizing French, and centralizing royal power (Grinberg
2006; Kibbee and Keller 2018). Shortly afterwards, during one of his well-attended
table talks in the 1530s or early 1540s, Martin Luther remarked that ‘Germany has
so many dialects that people within thirty miles do not understand each other’.³
Luther clearly linked dialectal variation with geographical distance; every thirty
² Antesignanus (1554: 12): ‘sermonis quandam proprietatem, qua distinguitur loquelæ varietas, quę
semper solet contingere inter diuersos eiusdem nationis tractus: cùm hi paulò aliter loquantur quàm
illi, ac cuique proprium quidpiam è natali solo sit insitum’.
³ Luther (1919: 511): ‘Germania tot habet dialectos, ut in triginta miliaribus homines se mutuo non
intelligant’.
.      97
miles there was a dialect so different that it could not be understood by a

non-native speaker. Luther was, by the way, not alone in recurring to measures
of length to underscore the immensity of dialectal diversity. Several other early
modern authors expressed regional variation within a language in terms of miles
or milestones. One can read another early example in the Greek grammar of the
Spanish humanist Francisco de Vergara (†1545), who took great pains to explain
Greek diversity by referring to his native Iberian context. In doing so, Vergara
remarked that ‘indeed the differences of some letters and syllables occur with us
every tenth milestone in whatever direction’.⁴
Still, it was only after 1550 that the spatial conception of dialect as opposed to
language became part of theorizing on the term dialectus. Even though the idea
was rarely made explicit in the remainder of the sixteenth century, its frequency
steeply increased throughout the seventeenth and especially the eighteenth centuries.
Indeed, formulations of the geographical interpretation became ever more
substantial from the turn of the seventeenth century onwards, in which case
Antesignanus’s and Gessner’s works were not infrequently quoted or para-
phrased. For instance, Andreas Hoius (1551–1635), a professor of Greek from
Bruges, cited part of Antesignanus’s definition in his Short dissertation on the
seats and colonies of the dialects of the Greek language of 1620, even though
Hoius did not do him the courtesy of mentioning him by name.⁵ The renowned
grammarian Justus Georg Schottel (1612–76), in turn, resorted to Gessner’s
Mithridates when defining dialectus in his extensive 1663 description of the
German language.⁶ Schottel did explicitly refer to Gessner, but he did not notice
that the definition he attributed to Gessner was actually the one Gessner himself
had quoted, in Latin, from Clement of Alexandria (see Chapter 4, Section 4.3).
It may be clear from the above that early modern scholars directed most
attention to the dialect pole when discussing the locality of linguistic entities.
Their insistence on the restricted geographical range of dialects implied that
language covered a vaster space. Not everyone associated language with a large
region, however. What is more, several eighteenth-century scholars closely linked
it with the capital of a state, thus suggesting that the geographical coverage of a
language was about as restricted as that of a dialect. For example, the unidentified
author of the Encyclopédie article ‘Patois’, perhaps to be identified with Louis de
Jaucourt (1704–79), claimed that ‘one only speaks the language in the capital’.⁷
⁴ Vergara (1537: 210): ‘Literaru[m] vero & syllabaru[m] quarunda[m] differe[n]tię decimo quoq[ue]
miliari quaqua versum nobis occurrunt’.
⁵ Hoius (1620: 95): ‘Sive, ut παχυμερέστερον explanemus, Sermonis quandam proprietatem, qua
loquendi varietas distinguitur, quę inter diversos eiusdem nationis tractus plerumque solet existere’. See
van der Aa (1867) and Sacré (1994) for biographical information, the latter providing the correct year
of Hoius’s death.
⁶ Schottel (1663: 150): ‘Oder wie Gesnerus in Mithrid. sagt: Dialectus est dictio peculiarem alicujus
loci notam seu characterem præ se ferens’.
⁷ Anon. (1765: 174): ‘On ne parle la langue que dans la capitale’.
98   
In eighteenth-century England, one grammarian argued that ‘[a] Language is

always esteemed to be spoke the purest in and near the Capital, as London,
Rome, Athens, Jerusalem’ (Bayly 1756: 13). In such cases, language was obviously
identified with good language as opposed to the corrupt dialects or, in France, the
so-called patois. Indeed, there, the concept of patois, closely cognate to dialect,
emerged during the seventeenth century. This word first denoted ‘way of speak-
ing/acting, speech’, later ‘natural language, mother tongue’, and eventually, from
the seventeenth century onwards, ‘local variety of a language, generally considered
corrupt’ (Thomas 1953: 93–9, 105). Apart from carrying a strongly pejorative
connotation, the term patois was associated with delimited areas, often provinces,
just as dialect was.⁸
7.2 Two topical topoi
Remarkably, the definition of the Greek philosopher Diogenes of Babylon, ana-

lysing diálektos as a linguistic expression with regional characteristics (Chapter 2,
Section 2.1), was nearly absent from early modern theorizing. Gessner’s teacher
Theodore Bibliander (1548: 107) was one of the few early modern scholars who
quoted it, but he left out, surprisingly from a modern point of view, the part in
which regional particularity was alluded to. Bibliander’s work is nevertheless of
importance for intuitively linking dialectus, a term he used rather ambiguously,
with a restricted geographical area. He stressed, on the one hand, that each city of
a nation had something distinct in its speech (Bibliander 1542: 54) and, on the
other hand, that individual provinces had their own dialects (Bibliander 1548:
19–20). In doing so, Bibliander provided early examples of two widespread topoi.
First, dialects were often associated with the regional-administrative entity
generally termed a province, or provincia in Latin, rather than with isogloss
lines, which is very much a modern concept. In late antique and medieval
writings, regional variation had already been connected with provincia, most
notably by the Early Christian author Jerome. He noticed that, although it was
the primeval tongue, Hebrew, too, was a diversified language, since ‘every prov-
ince and region had its properties and could not avoid the native sound of
speaking’.⁹ This link became widespread only during the early modern era,
however. The French Hellenist Henri Estienne (1528/31–98), for instance,
believed ‘French’ dialects such as Picard and Norman to be connected to the
provinces of France (1587: 93). Even the Ancient Greek dialects came to be
associated with provinces in a remarkably anachronistic fashion. In a Greek
⁸ See e.g. Besnier (1674: 23). For the pejorative connotations see Chapter 9.
⁹ Commentarii in evangelium Matthaei 1.1452–6: ‘unaquaeque provincia et regio habebat proprie-
tates suas et vernaculum loquendi sonum vitare non possit’.
.    99
grammar published in 1752 in Mannheim, for instance, ‘Greek dialect’ was

defined as ‘that mode of speaking which is different in the different provinces of
Greece, which uses one language’.¹⁰ Yet the Greek dialects never corresponded to
such administrative regions, since ancient Greece was rather a collection of city-
states than a united area with a centralized government.
Second, it was an early modern topos to state that language varied from city to
city and even within individual cities. This idea had already been expressed by
the Florentine author Dante, who claimed that there was great variation not only
across different Italian towns, but even within cities such as Bologna (De vulgari
eloquentia 1.19.3). It was repeated by several early modern scholars, most of them,
however, independently of the Tuscan poet. Examples are legion. The Italian
humanist historian Benedetto Varchi, for instance, remarked that a language
could show differences in pronunciation from city to city and even from village
to village. Varchi (1570: 89–90) compared this great diversity of speech with the
individuality of everyone’s handwriting. Well before Varchi, the great Roman
humanist-printer Aldus Manutius (c. 1449/51–1515) had asked himself: ‘What is
the reason that every city has a peculiar tongue, that very often also in the same
city they speak variously?’¹¹ Some scholars went even further, claiming that a
language could show differences from house to house or from family to family
(e.g. Trissino 1529: .ii). In sum, geographical entities of different sizes were
involved in discussions of dialectal variation. This diverging geographical scaling
of language not only suggests that early modern authors were aware of the
gradational nature of linguistic diversity, but is also in line with a main charac-
teristic of Renaissance conceptions of space in general: its fluidity. In fact, the
apparent intangibility of dialectal variation might have contributed to the early
modern sense of the ‘inherent instability’ of space, which was also connected to,
among other things, the lack of fixity of state borders (Kümin 2015: 230).
7.3 Regional language variation: a universal phenomenon?
And out of experience we see for display that all the languages of the
entire world are somewhat different from locality to locality by a
distance of no more than two milestones. If you would want to
proceed farther, you will find such diverse tongues that they barely
understand each other.¹²
¹⁰ Goldhagen (1752: 32): ‘Est autem Dialectus Græca modus ille loquendi in diversis Prouinciis
Græciæ unâ linguâ utentis diversus’.
¹¹ Manutius (1496: *.ii): ‘Quid, quod unaquaeq[ue] urbs peculiarem habet lingua[m]· plerunq[ue]
etiam in eodem oppido uarie loquuntur’.
¹² Haloinus (1978 [1533]: 55): ‘Experientiaque videmus ad oculum linguas omnes totius orbis de
loco ad locum non amplius duorum miliarium distantia aliquantulum diversas esse; si longius
procedere velis, tam diversas invenies, ut vix alius alium intelliget [sic]’.
100   
This observation by the Flemish nobleman and diplomat Georgius Haloinus

(c. 1470–1536/37) on linguistic diversity is one of the earliest explicit acknow-
ledgements of the universality of regional language variation, if not the earliest.
Haloinus formulated the idea in his Restauration of the Latin language, a work
published in 1533 but written already in 1508, in which he expressed his prefer-
ence for usage above grammar as the point of reference for correct language. His
many diplomatic travels for the Habsburg administration brought him into
contact with a wide range of different dialects and languages and heightened his
awareness of the phenomenon of regional linguistic diversity and its universality
(Matheeussen 1974: .15, .18, .172, .290; 1977: 310–11).
The idea was absent from ancient and medieval considerations on language,
most likely because Latin was widely regarded as a stable, unchangeable tongue.
It is, for instance, telling that the Italian humanist Niccolò Perotti still held around
1478 that ‘there is sometimes’—so not always—‘diversity of speech in one tongue’,
which he exemplified by referring to the Greek language (see Chapter 5,
Section 5.1). Haloinus, however, believed that there was variation within Latin
as well; this he motivated by pointing to the Patavinitas of the Roman historian
Livy (59 – 17), an alleged regional particularity in Livy’s speech reported by
Quintilian and claimed to be due to his Paduan roots (Van Rooy 2018a). The idea
that Latin was not as uniform as was generally assumed had its roots in the
Quattrocento, during which Italian humanists debated intensively about the
possibility of class-based linguistic variation in ancient Rome (Chapter 2,
Section 2.4). It was not until the early sixteenth century, however, that scholars
put forward the universality of dialectal diversity. This development was surely
related to the ongoing emergence, around that time, of the distinction between
language and dialect, which was the product of intuitively comparing different
contexts of regional diversity.
The idea of dialectal diversity as a universal phenomenon was most widespread
in Germanic-speaking territories. The renowned Bruges-born mathematician
Simon Stevin (1548/9–1620), for instance, emphasized the universality of dialectal
variation in his Discourse on the dignity of the Dutch tongue, written in Dutch and
prefixed to his Principles of statics: ‘It is common to every language that in one
region of the country it is spoken somewhat differently from another’.¹³ In the
seventeenth and eighteenth centuries, especially from about 1650 onwards, the
universality idea became a kind of refrain and even came to be presented as
generally accepted knowledge: ‘It is well known that not only divers nations, but
divers provinces or parts of the same nation who use the same language differ in
their dialect and manner of pronunciation’ (Poole 1683: ii.1). The universality of
¹³ Stevin (1586: b.1): ‘Tis yder spraeck ghemeen datse inden eenen oirt des landts wat anders
uytghesproken wort als op den anderen’.
.    ? 101
dialectal diversification led some authors to characterize it as an inevitable process,

linked to the general variability of human affairs (e.g. Howell 1642: 144).
7.4 Nation or nations? The ethnic conception

of the language/dialect distinction
‘Dialect’ . . . is speech peculiar to some people and this in the same

language.¹⁴
Georg Pasor (1570–1637), a German Calvinist professor of theology and classical

languages and one of the first specialists in New Testament Greek, associated the
Greek term diálektos in his work on the nature of Biblical Greek with the entity of
‘people’. This ethnic component in Pasor’s definition comes as no surprise, since
there was a close link in Greek scholarship between diálektos and éthnos, ‘nation,
tribe’, probably an even stronger one than there was between diálektos and tópos,
‘place’. In line with Greek tradition, Pasor and others connected dialect with a
nation, gens, natio, or populus in Latin, in the non-political sense of ‘group of
people sharing the same ethnic background and living in the same area’. This
coupling of dialect with nation could be taken to imply that language covered
several nations, yet such an idea was rarely made explicit.
Already in the early sixteenth century, diálektos and dialectus were frequently
explained as ‘speech particular to a certain people’. In the majority of these
instances, however, it is impossible to posit a subsumption of dialect under lan-
guage. For example, in a basic Greek–Latin lexicon of 1514, attached to an edition of
the Greek and Latin New Testament which appeared in Alcalá de Henares in Spain,
diálektos was explained as ‘idiom, tongue, speech peculiar to whatever nation’.¹⁵ The
compiler of the word list cannot, however, be claimed to have presupposed a
language/dialect contrast, as he took diálektos to refer very generally to nation-
particular speech, without opposing it to a language-like entity.
French Hellenists active around the mid-sixteenth century seem to have already
associated the entity of ‘nation’ with the conceptual pair. Petrus Antesignanus
(1554: 11), for instance, emphasized that there were as many dialects as there were
Greek nations, all of which used the Greek language with regional differences.
Elsewhere, however, Antesignanus stated that dialectal variation ‘usually occurs
between different regions of the same nation’, a contradiction he failed to resolve
(see Section 7.1 for the full quote). The association of dialect with nation was also
implicit in a German grammarian’s classification of the dialects of his native
tongue into two principal groups:
¹⁴ Pasor (1632: 1): ‘Διάλεκτος . . . est sermo cuique populo peculiaris, idque in eadem lingua’.
¹⁵ Guillén de Brocar (1514: b.iii–iv): ‘Idioma.Ligua [sic].sermo proprius cuiusq[ue] nationis’.
102   
just as the Germans are commonly and abundantly divided into two nations
today, the High and Low Germans, likewise our language is generally divided
into two particular idioms: High German and Low or Saxon German.¹⁶
The most interesting sixteenth-century testimony for the relationship of a

language to nation is provided by the Italian humanist Benedetto Varchi.¹⁷ In
his dialogue L’Hercolano, Varchi defined lingua as ‘nothing else than the speech of
one or more peoples (popoli), which employ, in expressing their concepts, the
same words in the same meanings and with the same accidents’.¹⁸ In what
followed, he elaborated on each of the elements included in his definition and
by consequence also on the phrase ‘of one or more peoples’. On the one hand, he
introduced the element ‘of one people’, because the speech of a group of friends
could also have particularities of its own. This specificity was not a reason to call it
a lingua, however, as it should rather be called gergo, ‘slang, jargon’. On the other
hand, it was possible that one and the same lingua was used by several peoples, as
in the cases of Latin and Slavic, either as a result of natural developments or by
accident. So, in the eyes of Varchi, who explicitly subsumed dialect under language
when discussing the Greek dialects elsewhere in his work, a speech form was a
language as soon as it was spoken by at least one people. He did not, however,
explain what the ethnic coverage of a dialect was, according to him.
The generic meaning ‘speech form particular to a certain nation’ of the term
dialectus increasingly interfered with interpretations of the language/dialect pair
from the end of the sixteenth century onwards. As this sense of the word was
inherited from Greek definitions of diálektos, it is not surprising that one encoun-
ters the first clear indications of this interference, prefigured in the thought of
Antesignanus, in handbooks for the Greek language and its dialects. For instance,
the English antiquarian and Hellenist William Camden (1551–1623) stated the
following in his 1595 Greek grammar:
One generally reckons four dialects of the Greek language or principal forms of
speaking according to the number of the principal nations of Greece, which vary
from the common language in some respects, i.e. Attic, Ionic, Doric, Aeolic.¹⁹
¹⁶ Albertus (1573: .7): ‘Sicut Germani communiter et largè in duas gentes hodie diuiduntur,
Superiores & Inferiores. Ita generaliter diuiditur lingua nostra in duo Idiomata, in Oberländisch/ vnnd
Niderlendisch/oder Sächsisch Teutsch‘.
¹⁷ On Varchi’s linguistic thought and classification in general see Marazzini (1993: 267–73).
¹⁸ Varchi (1570: 87): ‘L, ò  ,  ̀ ,    ’, ò ̀
,  , ò  ,     ,    
 ,  ’  ’.
¹⁹ Camden (1595: .1): ‘Qatuor numerantur Græcæ linguæ Dialecti siue loquendi formæ
præcipuæ pro numero præcipuarum Græciæ gentium, quæ à linguâ communi in nonnullis variant,
nimirum Attica, Ionica, Dorica, Æolica’.
.    ? 103
The ethnic interpretation of the conceptual pair soon appeared outside Greek
handbooks as well. An intriguing case in point is the Flemish theologian and
orientalist Johannes Drusius (1550–1616), who observed that dialectus was a term
with different meanings and devoted particular attention to its ethnic conception.
Drusius was led to do so by the Hebrew translation of a New Testament phrase
containing the Greek term diálektos. The Hebrew word lāšṓn (‫)ָלשׁוֹן‬, ‘tongue,
language’, was used to render diálektos, and Drusius took the Hebrew term to
be equivalent to Latin lingua, ‘language’. This apparent mismatch between
diálektos, ‘dialect’, and lāšṓn, ‘language’, called for some explanation, he must
have thought. Drusius pointed out that dialectus had two principal meanings: the
‘wider’ interpretation as ‘speech particular to a people’, which was distinct from its
‘proper’ sense of ‘speech subsumed under a language’. He clarified the latter
meaning of dialectus as ‘a peculiar manner of speaking in an otherwise identical
language’ and further explained the former sense as follows: ‘distinctiveness of
speaking of different peoples, whether they are of the same language or not’.²⁰ In
other words, the broader ethnic meaning did not presuppose a subsumption of
dialect under language, whereas the stricter interpretation did. Since the word
dialectus remained polysemous throughout the early modern era, it should come
as no surprise that many scholars employed it in both of these different but related
meanings, often in highly equivocal ways.
Not infrequently the ethnic meaning was irrefutably present, with a language/
dialect distinction being assumed at most, as, for instance, in the work of Claude
de Saumaise (1588–1653), a French philologist and professor at Leiden university
(Figure 7.1). Saumaise even overemphasized the ethnic parameter in defining
diálektos, since he wanted to refute the existence of a so-called Hellenistic Greek
dialect in antiquity. He did so in his fierce controversy with his great rival Daniel
Heinsius (1580–1655), a humanist from Ghent who had presupposed the exist-
ence of this Hellenistic dialect.²¹ To speak of a Hellenistic dialect, Saumaise
claimed, a Hellenistic tribe was needed, since Greek scholars had defined
diálektos as ‘speech particular to a certain tribe’. As he followed Greek tradition
rather closely, he did not subsume dialect under language in his own definitions.
Saumaise (1643a: especially 6–7, 459–61; 1643b: 312) defined diálektos in the first
place as a living, nation-specific variety, not necessarily subsumed under a lan-
guage and not necessarily related to another diálektos genealogically (Considine
2010: 89). Or in his own words: a diálektos was a ‘distinctive ethnic and local mark’
²⁰ Drusius (1616: 76): ‘Est autem dialectus proprie in eadem alioqui lingua peculiaris loquendi
modus, veluti apud Græcos cum una sit lingua quinque tamen sunt dialecti. Sed latiore significatione
sumitur Act. 2. 8. & alibi. Sic dico, suus & proprius cujusque gentis sermo .i. diversorum populorum
discrimen loquendi sive sint ejusdem linguæ sive non, vocatur διάλεκτος’.
²¹ For this much-studied controversy see e.g. de Jonge (1980: 32–4); Muller (1984b: 391–2);
Considine (2010).
104   
Figure 7.1 Claude de Saumaise (1588–1653)

in speech.²² It is, however, hard to deny that Saumaise assumed the existence of a
language/dialect distinction elsewhere in his work, as, for instance, in the follow-
ing statement: ‘The word glôssa seems more general, diálektos more specific about
a “derivation” of whatever language’.²³ He moreover employed the expression non
nisi dialecto differre, ‘to not differ except in terms of dialect’, which indicates that
he associated dialect with accidental variation and language with substantial
differences (Saumaise 1643a: 356, 368; Chapter 8, Section 8.1). In short, even
though Saumaise doubtlessly presupposed a distinction between language and
dialect, the relationship of the ethnic conception of dialect to his understanding of
the conceptual pair remained unclear. This ambiguity must be viewed against the
background of his scholarly modus operandi: chaotic and repetitive.
²² Saumaise (1643a: 61): ‘χαρακτὴρ ἐθνικὸς καὶ τοπικός’.

²³ Saumaise (1643a: 456): ‘verbum γλῶσσα generalius videtur, διάλεκτος magis speciale de
παραγωγῇ cujusque linguæ’.
.      ‘’? 105
The insistence on the interpretation of the conceptual pair in terms of ethnic

coverage manifested itself differently in the thought of several of Saumaise’s
contemporaries. They conceived of a dialect as a variety of a language that was
the usual speech form native to a specific nation, whereas one needed to study to
master a language. For example, the German grammarian Justus Georg Schottel
remarked the following when elaborating on the distinction between everyday
linguistic usage, covering also dialectal differences, and the German language:
everyday usage is instilled from the cradle and acquired through itself, but the
language is learned through nothing else than artificial instruction and the
necessary zeal and reflection.²⁴
As a consequence, a common language was sometimes seen as ethnically

unmarked. The German diplomat and scholar Caspar Schoppe (1576–1649), for
instance, declared:
if someone who otherwise recognizes all dialects, would hear someone else
speaking in that [sc. the common German language], he would certainly approve
of the speech itself, but would not be able to determine to what nation
he belongs.²⁵
Schoppe added that the German common language was easily learned in Speyer, a
city in the south-east of modern Germany, and at the imperial court, where people
from all over Germany flooded together and dialectal elements were avoided in
order not to be exposed to ridicule. In other words, dialect was bound to a nation,
whereas language was neutral in terms of nation, an interpretation miles away
from modern politicized conceptions of the distinction.
7.5 Towards a political interpretation of ‘nation’?
From the mid-seventeenth century onwards, language was increasingly associated

with the entity of the state and its inhabitants, sometimes termed nation, which
began to receive a more politically coloured meaning. Indeed, language became
linked more closely to state structures and was increasingly seen as an important
symbol of the distinctive character of a particular state. This politicization of
²⁴ Schottel (1663: 168): ‘Weil der altages Gebrauch von wiegen an eingeflösset/und durch sich selbst
angenommen; Die Sprache aber/mit nichten anders/als durch kunstmessige Anleitung un[d] erforder-
ten Fleiß und Nachsinnen/erlernet wird’.
²⁵ Schoppe (1636: 51): ‘si quis ea loquentem audiat, qui ceteroqui omnes Dialectos agnoscit,
sermonem quidem ipsum probet, nequaquam tamen, cuius ille nationis sit, constituere possit’.
106   
language seems to have happened first in England and later on in France, too, and
is reflected in linguistic works of the time. For instance, in the preface to his
dictionary The new world of English words, Edward Phillips (1630–c. 1696)
defined dialect as follows:
a Dialect is but the self same Language, spoken in several Provinces of the same
Nation, with some small difference; as the pronouncing of a vowel either broader
or finer, or some little variation of a word or syllable. (Phillips 1658: b.3)
Here, Phillips’s association of language with nation was no doubt inspired by his
native English context, both linguistic and political. He referred to the kingdom of
England and its provinces, which were tied, respectively, to the English common
language and its dialects.
In a similar vein, the lemma for nation in the first edition of the dictionary by
the Académie française, the French Academy, assumed the existence of one and
the same language for a nation:
N. . . . All inhabitants of one and the same state, one and the same country,
who live under the same laws and employ the same language etc.²⁶
This development no doubt related to the advancing linguistic standardization

and the ever-increasing political centralization occurring in Western European
sovereign states such as England and France from the sixteenth century onwards
(e.g. Trudeau 1992; Rice and Grafton 1994: 111–24). More generally, the associ-
ation of vernacular languages with nations was a product of the Renaissance,
during which political borders became increasingly determined by linguistic
factors and humanists started to feel the need to distinguish their own nation
clearly from others (Hirschi 2012: especially 104–79; Chapter 6, Section 6.2).
However, in the majority of the early modern source texts, the ethnic inter-
pretation of the language/dialect distinction, with a language covering several
nations in the non-political sense and a dialect only one, is more prominent
than the association of language with nation as inhabitants of a political state
and of dialect with the provinces of such a state. One might therefore even assert,
albeit cautiously, that the emphasis on the ethnic interpretation of the conceptual
pair deconstructed to a certain extent the national identities invented by humanist
scholars, especially since speakers of a certain dialect were very often clearly
demarcated as a separate linguistic group by early modern scholars.
²⁶ Académie française (1694: .110): ‘N. . . . Tous les habitants d’un mesme Estat, d’un mesme
pays, qui vivent sous mesmes loix, & usent de mesme langage &c.’.
.   :  ? 107
7.6 Space and nation: diagnostic criteria?
The interpretation of the conceptual pair in terms of ethnic coverage was closely
associated with the spatial conception of dialect, discussed earlier in this chapter.
Indeed, a nation was intrinsically tied to a certain locality in the eyes of many early
modern scholars defining dialect. Yet, although present in many definitions, the
geographical and ethnic conceptions of the language/dialect distinction were
usually not put into practice as diagnostic criteria, that is to say, as litmus tests
to determine whether a specific linguistic variety was a dialect or a language. What
is more, in cases where these interpretations were employed to determine whether
a linguistic variety deserved a certain label, it was not to argue whether a speech
form was either a language or a dialect but rather to deny dialect status to it.
I have already mentioned how Claude de Saumaise relied on the ethnic
meaning of diálektos to refute the existence of a Hellenistic dialect; there was no
Hellenistic nation or land, so there cannot have been a Hellenistic dialect of the
Greek language. Let me provide an additional example to elucidate my point.
Certain early modern Hellenists had posited the existence of a dialect particular to
Greek poets. Soon several of their colleagues invoked the locality of dialect or its
belonging to a specific tribe to challenge the existence of this Greek poetical
dialect. Two eighteenth-century manuals for the Greek language offer excellent
examples of this usage. In a handbook for the Greek dialects published in
Nuremberg in 1782, one can read the following:
Besides, the poetic dialect, usually added to these four dialects, cannot be called a
dialect properly speaking, as it is rather a certain kind of speech not particular to
a nation, but to a certain class of writers only.²⁷
A similar statement is found in a Greek grammar published two years earlier in

Leipzig, which rebutted the existence of the poetical dialect by means of the
geographical interpretation of dialectus: ‘Some turn the liberty the poets take in
their verses into a dialect and call it poetical dialect. Such a dialect, however,
presupposes a poetical city or region’.²⁸ Even though in these cases the spatial and
ethnic conceptions of dialect were applied as touchstones in determining whether
a linguistic variety was a dialect, this decision occurred without putting the
subsumption of dialect under language into practice.
²⁷ Facius (1782: ): ‘Quae praeterea his quatuor Dialectis vulgo additur Poetica, proprie Dialectus
dici nequit, cum dictionis potius sit quoddam genus, non genti, sed scriptorum tantum ordini cuidam
proprium’.
²⁸ Haas (1780: 67): ‘Manche machen die Freyheit, deren sich die Poeten in ihren Versen bedienen,
zu einem Dialekt, und nennen ihn dialectum poët[icam]. Ein solcher Dialekt aber setzet eine poetische
Stadt oder Landschaft voraus’.
108   
7.7 Conclusion
Greek definitions of the term diálektos left their mark on early modern concep-
tions of the language/dialect opposition in two principal ways. First, from about
1550 onwards, dialect was increasingly understood as a variety restricted in terms
of space vis-à-vis the language subsuming it. This geographic conception probably
resulted from the availability in print of Clement of Alexandria’s Stromata, which
contained several important definitions of diálektos. The geographical interpret-
ation of the conceptual pair was, however, not actively used to distinguish a dialect
from a language, unlike in modern linguistics (see e.g. Hinskens et al. 2005: 1).
Still, it remains a fact that the geographical interpretation of the conceptual pair
frequently occurred in definitions of the term dialectus.
Second, a dialect was claimed to be particular to a certain tribe or nation,
implying that a language covered several different tribes or nations. This ethnic
interpretation of the conceptual pair became widespread only in the seventeenth
century, well after the geographical meaning, and it is unclear why this delay
occurred. Before that, the ethnic interpretation of the term diálektos/dialectus,
already present in Greek scholarship, barely interacted with the, at that time still
loose, semantics of the conceptual pair. Just like geographical coverage, the ethnic
conception was not actively and consciously used as a criterion in determining
whether a linguistic variety was a language or a dialect.
In short, even though the majority of early modern scholars assumed that
dialects had a limited geographical and ethnic coverage as opposed to a language,
in practice they relied on other criteria to determine the language or dialect status
of a linguistic entity.
8
Aristotle’s legacy
Substance, accidents, and mutual intelligibility
The impact of the Greek tradition is noticeable not only in the spatial and ethnic
conceptions of dialect as subsumed under language. More indirectly, it can be
perceived in other emergent interpretations of the conceptual pair, too, for
instance in the following formulation by the renowned Swedish poet and philolo-
gist Georg Stiernhielm (1598–1672), founded on the notions of substance and
accidents from Aristotelian philosophy: ‘Languages differ among each other in
substance, in their foundation as it were, but dialects differ in accident’.¹ The
present chapter investigates the widespread early modern idea that distinct
languages exhibited substantial differences, whereas related dialects were only
superficially variegated, and its consequences.
8.1 Extending the ‘to differ only in dialect’ phrase
The substance/accidents interpretation was already prefigured in the thought of

Roger Bacon (see Chapter 3) and was implicit in the phrase ‘to differ only in
dialect’, first coined in the 1540s (see Chapter 5, Section 5.5). It was, however, only
in the course of the seventeenth century that it was explicitly formulated as a
diagnostic criterion. Before looking into this evolution, however, I should briefly
consider its philosophical background. There is no doubt that the traditional
distinction of Aristotelian ontology between substance in the sense of ‘fundamen-
tal essence’ and accidents as ‘non-fundamental properties’ constituted the intel-
lectual backdrop for this interpretation of the language/dialect distinction. This
conception of it may therefore well be dubbed the ‘Aristotelian criterion’. It must
be understood against the background of the prevailing Aristotelianism in the
arts curricula of many universities in the first half of the early modern period
(Schmitt 1983; Brockliss 2003). Aristotle (Figure 8.1) was, however, never expli-
citly mentioned by scholars discussing the criterion, which is surely due to the
fact that the substance/accidence opposition had become received knowledge.
This silence resembles a process called, in the sociology of science, ‘obliteration
¹ Stiernhielm (1671: c.4): ‘Linguæ inter se substantia, ceu subjecto; Dialecti vero Accidenti differunt’.
DOI: 10.1093/oso/9780198845713.001.0001
110  ’  
Figure 8.1 Bust of Aristotle, after an original of about 330 

by incorporation’, during which ideas become so established that the author first
expressing them is no longer cited (Merton 1968: 27–8, 35).
How did the Aristotelian criterion become explicit? To answer this question,
I have to take a start in the early seventeenth-century Low Countries, where there
was a great upsurge in interest in linguistic diversity (see especially Van Hal 2010a).
In 1605, the Leiden history professor Paulus Merula (1558–1607) drew attention to
the following research lacuna concerning the ancient tongues of Gaul:
What that difference in language was, whether it was in dialects only or in the
entire language, so that the Aquitanians, Celts, and Belgians each had a different
language, has not yet been distinctly uncovered by anyone, at least as far as I know.²
² Merula (1605: 419): ‘Quod in Lingua discrimen illud fuerit, an in Dialectis solum, an in tota
Lingua, ut ea alia fuerit Aquitanis, alia Celtis, alia Belgis, nemo adhuc, quod ego quidem sciam, diffinite
aperuit’. See also Chapter 10, Section 10.3.
.   ‘    ’  111
Here, the phrase ‘to differ only in dialect’ was extended so as to include the
conceptual counterpart of language, too. What is more, Merula further deter-
mined the term lingua by means of the Latin adjective totus, meaning ‘whole,
entire’, clearly contrasting dialect to language. Extending the phrase in this
manner was principally a seventeenth-century development, with the Low
Countries as its focal point. Some forerunners of this evolution can, however, be
found in the late sixteenth century. The Scottish humanist historian George
Buchanan (1506–82), for instance, ‘believe[d] that the Scots did not differ from
the Britons in the entire speech but rather in dialect’ in ancient times.³ A little
further Buchanan expressed his astonishment that
even now after so much time, with the languages of neighbouring nations
adulterated and changed for a large part by the arrival of so very many peoples,
the Brits still differ among each other not so much in the entire speech as in
particularity and dialect.⁴
Such expansions of the phrase ‘to differ only in dialect’ are a sign that scholars
were moving towards an explicit formulation of the Aristotelian criterion. The
Dutch pastor Abraham Mylius (1563–1637), the first early modern scholar to
systematically treat the themes of linguistic diversity, change, and kinship, in his
book Belgian language, provided an even clearer example of this tendency.
Consider, for instance, the following statements of his:
Even though indeed the Chaldean language, by the progression of time, has
considerably changed in dialect from Hebrew, yet, as far as substance and root
are concerned, it has remained the same.⁵
It can happen that that first distinction of languages [sc. the confusion of tongues
at Babel] did not so much occur in respect to the substance of words as, for the
most part, in respect to dialect and pronunciation.⁶
As a matter of fact, even if the words display some variation of dialect from our
tongue, yet the substance is still the same, except in a few words, which time has
worn out, as it has happened in the Greek and Latin language, and as it happens
in all languages.⁷
³ Buchanan (1582: 19): ‘Sed nec Scotos a Brittonibus toto sermone, sed dialecto potius discrepasse
arbitror’.
⁴ Buchanan (1582: 20): ‘nu[n]c etiam tanto post tempore tot gentium aduentu adulteratis, & magna
ex parte mutatis vicinarum nationum linguis adhuc Britanni non tam sermone toto, quam proprietate,
& dialecto inter se discrepe[n]t’.
⁵ Mylius (1612: 85): ‘Etiamsi vero Chaldæa lingua, progressu temporis, dialecto non parum sit ab
Hebræa mutata, tamen, quod ad substantiam & radicem attinet, mansit eadem’.
⁶ Mylius (1612: 85): ‘Fieri posse, ut illa prima discriminatio linguarum non tam circa dictionum
substantiam, quam ut plurimum circa dialectum pronuntiationemque facta sit’.
⁷ Mylius (1612: 151): ‘Nam etsi in vocibus dialecti quædam sit varietas a nostra, tamen substantia est
adhuc eadem, nisi in paucis, quas ævum detrivit: ut factum fuit in lingua Græca & Latina, & fit in
omnibus linguis’.
112  ’  
For these views, Mylius was inspired by the close kinship among the Ancient
Greek dialects (cf. Mylius 1612: 90, 152). Put differently, even though the
Aristotelian criterion was an early modern innovation, it was nevertheless inspired
by Greek philosophical categories and sometimes corroborated by the prototyp-
ical example of Greek linguistic diversity.
Aristotelian categories were, however, not the only means used to express this
conception of the language/dialect distinction. To refer to the substantial nature of
language-level differences, some scholars used the image of ‘foundation’ (fundus),
whereas others claimed that the various dialects of one language agreed in the
‘root’ or ‘radix’.⁸ Family relationships were occasionally also used to stress the
superficial differences among dialects (e.g. Saumaise 1643a: 19–20).
Towards the middle of the seventeenth century, the Aristotelian criterion
became part of more general reflections on the conceptual pair. The extravagant
Berlin-born orientalist Christian Ravis (1613–1677) offered the following analysis
of closely related language varieties in his Sketch of Hebrew orthography and
analogy of 1646, published in Amsterdam:
Languages that are generally considered different yet connected through all essen-
tial elements, even though they apparently are very distinct by some and indeed
many various and innumerable accidental elements, are nevertheless dialects to
each other, and they constitute an integral language and a perfect body.⁹
In stating this idea, Ravis was one of the earliest scholars, if not the earliest, to
formulate the criterion in such explicit terms. After him, it became more common
to do so. Not unsurprisingly, it first occurred in the work of scholars active in
England, where another major linguistic work of Ravis, his Generall grammer for
the ready attaining of the Ebrew, Samaritan, Calde, Syriac, Arabic, and the Ethiopic
languages, was printed in 1650. In this book, one could read similar statements
(e.g. Ravis 1650: 44–5, 156). In the 1656 dictionary of Thomas Blount (1618–79),
for instance, dialect was defined as follows:
Dialect (dialectus) is a manner of speech peculiar to some part of a Country or

people, and differing from the manner used by other parts or people, yet all using
the same Radical Language, for the main or substance of it.
(Blount 1656: s.v. dialect)
The Aristotelian criterion soon also appeared outside England in the work of
Georg Stiernhielm, who had met Ravis in Sweden not long after the publication of
⁸ For ‘foundation’ see e.g. Schrieckius (1615: a.4). For ‘radix’ see e.g. Shelford (1635: 239).
⁹ Ravis (1646: 42): ‘Linguæ vulgò æstimatæ diversæ, per omnia tamen essentialia junctæ, licet
aliquibus & quidem multis varijsque & innumeris accidentalibus apparenter distinctissimæ, sibi
invicem tamen dialecti sunt, & linguam aliquam constituunt integram, corpusque perfectum’.
. :    ? 113
the latter’s Sketch (see Chapter 12). However, even after the mid-seventeenth
century, the criterion was usually taken for granted. Indeed, relatively few authors
included it in their definitions. The primarily implicit nature of the criterion is
in line with the fact that early modern scholars assumed the conceptual pair to
be an obvious given, rather than an innovation of their own, as I have argued in
Chapter 6.
8.2 Mithridates: polyglot or not?
The Aristotelian criterion was eagerly put into practice by early modern scholars.
I have discussed in Chapter 5 (Section 5.5) how the coining of the phrase ‘to differ
only in dialect’ was related to the sixteenth-century interest in the poorly docu-
mented languages of ancient Gaul and Germany and their interrelationships. Let
me consider here another discursive use, or rather abuse, of the criterion.
Overcoming extant linguistic diversity was a difficult enterprise for early modern
scholars. They resorted to the uniform Latin tongue, the common and widely
understood scientific language, as an antidote to this historical reality. Another,
later solution consisted in attempts at designing a universal language that would
make other languages obsolete (e.g. Wilkins 1668). A more difficult and time-
consuming method was learning different languages. Indeed, a polyglot compe-
tence was a source of admiration, whereas a lack of it was often held up to
mockery, certainly in philological circles. For instance, the alleged ability of
Mithridates , ancient king of Pontus, to converse with the twenty-two tribes
he ruled in their native tongues was widely known and applauded.¹⁰ Perhaps the
best testimony to this appraisal is the fact that Conrad Gessner named his
language catalogue after him. However, the language/dialect distinction and its
interpretation in terms of substance and accidents made it possible to detract from
a polyglot’s competence by emphasizing the alleged dialectal relationship among
the tongues mastered. In fact, the nobleman Georgius Haloinus had already
suggested in the early sixteenth century that king Mithridates either only knew
some phrases of those twenty-two languages or spoke languages that ‘were
neighbouring and quite similar, as experience teaches that this occurs in almost
all regions’.¹¹ Similarly, the historian Paulus Merula considered it unlikely that the
Pontic king spoke twenty-two ‘original languages’ (linguae primigeniae), implying
that he had mastered different dialects of a restricted number of original lan-
guages. Merula (1605: 209) argued in this context that speaking different varieties
of Germanic or Romance did not grant one the title of polyglot, multi-linguis in
¹⁰ See Pliny the Elder, Naturalis historia 7.88, for the classical source of this information.
¹¹ Haloinus (1978 [1533]: 126): ‘linguae illae vicinae et fere similes erant, ut in omnibus fere
regionibus fieri experientia docet’.
114  ’  
his Latin, either. This idea was picked up by a number of later scholars. The
Anglo-Welsh literary author and grammarian James Howell (c. 1594–1666), for
instance, even expressed it by means of a variation on the ‘to differ only in dialect’
phrase:
There is also true Greek spoken in some parts of the lesser Asia, where there is no
place upon the surface of the earth, for the proportion, where so many differing
Languages are spoken, yet most of them are but Dialects and subdialects; so that
of those two and twenty tongues, which Mithridates is recorded to have under-
stood, above two parts of three, I beleeve, were but dialects. (Howell 1642: 150)
To sum up, the dialect concept and especially its Aristotelian interpretation was
put into practice as a discursive strategy to relativize a polyglot’s linguistic
competence, in particular that of the ancient king Mithridates.
8.3 A question of gradation: devising different levels of dialects
The Aristotelian criterion suggests that early modern scholars conceived of the
language/dialect distinction as a binary feature. However, many of them also
acknowledged that linguistic variation in general and dialectal variation in
particular come in different degrees. As a result, there was a tension between the
Aristotelian criterion and the realization that language diversity is a gradational
property. Was this mismatch resolved? And if so, how?
Even though there was a broad consensus that dialectal variation comes in
different degrees, no early modern scholar explicitly suggested that it constitutes a
continuum, let alone that he drew conclusions from it that had consequences for
the conceptualization of the language/dialect pair; its arbitrary nature was not
recognized. Scholars of the period did, however, usually assume that among
related dialects the similarities outweighed the differences. For this reason, some
of them pointed out that dialects could be designated by a common name, usually
that of the language of which they were dialects. The following was, for instance,
claimed for the poorly documented linguistic situation of ancient Gaul, for which
the Greek language served as a model: ‘nothing keeps us from giving it one and the
same common designation, even though it is not entirely the same, but slightly
varied, which also happened in the Greek language’.¹² Conversely, Georgius
Haloinus (1978[1533]: 126) remarked that even though French was one language,
it did have specific designations for its varieties, because they differed in certain
¹² Glarean (1538: 19): ‘Nihil tamen impedit quin communi appellatione eadem ac una nominetur,
quanquam non usquequaq[ue] eade[m], sed paululu[m] uariata, Id quod in Græca quoq[ue] lingua
accidit’. Cf. Chapter 4, Section 4.4.
.     115
respects. To the mind of some authors, dialect-level differences justified the

interchangeable usage of names: ‘the Chaldean and Syriac language are not
different languages but only different dialects, which is why someone who speaks
Chaldean is also said to speak Syriac’.¹³
But let me return to the acknowledgement that dialectal variation is a matter of
gradation: how did it manifest itself in early modern linguistic thought? Certain
scholars appear to have dissected the phenomenon of language variation in greater
detail by devising various levels of dialects. The German diplomat Caspar
Schoppe, for instance, distinguished as follows between two kinds of dialects in
a 1636 pedagogical work:
It seems proper to call some dialects “principal” or “general” and others “subject
to the principal dialects” or “specific”, which even though they depart from some
principal dialect in many respects are nevertheless related to it in far more
respects.¹⁴
Schoppe made the distinction between ‘general’ and ‘specific’ dialects when
emphasizing the importance of learning correct German, a study hampered by
the existence of the great variety of dialects. In what followed, he exemplified the
class of principal dialects by referring to the German of the Saxon city of Meissen,
the Greek of Attica, the Italian of Florence, the French of Orléans, and the Spanish
of Toledo, all perceived as the selected standard varieties of the respective lan-
guages. ‘Specific dialects’ of, for instance, the Meissen variety included Thuringian
and Franconian German. Schoppe might have drawn inspiration from the early
modern tradition of Greek dialect studies, where the principal literary dialects,
mostly Attic, Ionic, Doric, and Aeolic, were distinguished from and superposed
on secondary dialects (see Van Rooy 2020a: 19–24). Schoppe’s distinction was,
however, not predominantly inspired by literary motivations, as with the Greek
dialects, but by pedagogical and normative considerations of learning the correct
standard form of a language.
A similar differentiation of kinds of dialects can be found in Claude de
Saumaise’s 1643 Commentary on the Hellenistic tongue. Saumaise, however,
operated from a more strictly philological point of view, since he tried to analyse
and classify linguistic diversity by means of a tripartite conceptual hierarchy, with
a focus on the history of Ancient Greek. Saumaise (1643a: e.g. 154–6, 458–61;
1643b: 248) repeatedly differentiated between ‘generic dialects’, ‘specific dialects’,
and ‘highly specific dialects’, the last two being also labelled ‘local dialects’. In an
¹³ Leusden (1656: 201): ‘Lingua Chaldaica & Syriaca non sunt diversæ linguæ, sed tantummodo
diversæ dialecti; ideoque qui loquitur Chaldaicè etiam Syriacè loqui dicitur’.
¹⁴ Schoppe (1636: 46): ‘Dialectos alias vocare Principes placet, siue generales, alias principibus
subiectas siue speciales, quæ quamuis in multis à principe aliqua Dialecto recedant, in multo pluribus
tamen ad eam referu[n]tur’.
116  ’  
academic dissertation presented in Wittenberg on 9 February 1709, which like

Saumaise concentrated on the historical status of the Ancient Greek dialects, a
different three-layered hierarchy of dialects was proposed (Thryllitsch 1709:
.2–4; cf. Van Rooy 2016a: 470). The Ancient Greek language was divided
into four principal and primary dialects, each spoken by an entire tribe (dialecti
primariae, principales, or ethnikaí [ἐθνικαί]): Ionic, Attic, Doric, and Aeolic. Every
primary dialect comprised several secondary, regional dialects (dialecti secundar-
iae or egkhōŕ ioi [ἐγχώριοι]), which emerged as a result of the dispersion of the four
principal tribes. Finally, each secondary dialect comprised several city or local
dialects (dialecti urbicae or topikaí [τοπικαί]).
In brief, it became customary from the seventeenth century onwards to divide
the phenomenon of dialectal variation into two and sometimes even three hier-
archical levels. This division occurred primarily out of a desire to create more
order in the world’s linguistic diversity, even though pedagogical considerations
could also play a role, as in the case of Caspar Schoppe. Such distinctions had their
roots in early modern discourse on the Ancient Greek dialects, divided into
principal and minor dialects from the early sixteenth century onwards. What is
more, often Greek terminology was used in this context, even when it did not
concern Greek dialects, or at least not exclusively. In the course of the eighteenth
century, the distinction between primary and secondary dialects became ever
more detached from the Greek context and was increasingly applied to other
languages. It is, for instance, found in descriptions of diversity in Germanic (ten
Kate 1723: .243), Basque (Larramendi 1729: 12), and among the Oriental tongues
(Kals 1752: 58). There was great terminological proliferation to express different
kinds of dialects. For instance, the Dutch orientalist Jan Willem Kals (1700–81)
distinguished between ‘closer dialects’ (dialecti propinquiores) and ‘remoter dialects’
(dialecti remotiores; 1752: 58). One eighteenth-century Italian Hellenist even
coined a new word to denote lower-level dialects: the diminutive dialettuzzoli
(Mazzarella-Farao 1779: .). The distinction between primary and secondary
dialects was especially prominent in the thought of eighteenth-century German
scholars, who often referred to the former as Hauptdialekte or Hauptmundarten
and to the latter as, for instance, Nebendialekte.
The concept of secondary dialect generally entailed a value judgement. Indeed,
such dialects were often suggested to be somehow inferior and subordinate to a
primary dialect. To rid himself of such a connotation, a scholar could rely on the
more neutral term subdialect(us), coined around the middle of the seventeenth
century to denote a ‘variety subsumed under a dialect’. The concept had clear
forerunners in earlier texts, but it was not until the mid-seventeenth century that it
was condensed into a more or less demarcated notion, connected to a specific
term. The term was invented, apparently independently, by two different writers
active in the decades around 1650: the writer James Howell, who coined English
subdialect, and the Swedish philologist Georg Stiernhielm, who introduced Latin
.   117
subdialectus (Van Rooy 2019: 760). Lacking an intrinsically subjective semantics,

subdialect(us) was frequently used to simply denote a variety that was subsumed
under a higher-level dialect or genetically derived from a dialect. It thus had a
mainly classificatory function and constituted a more objective means to get a grip
of the layered and gradational nature of dialectal diversity. James Howell, for
instance, pointed out in his 1642 travel guide that someone journeying to France
should be aware of the great variation within French. This language had three
dialects, including Provençal, ‘whereof the Gascon is a subdialect’ (Howell 1642:
124). Howell’s analysis was clearly informed by a political bias; Walloon and
Provençal were subsumed under French, the language of the king and his
administration. Still, he did not want to make a prescriptive statement. Instead,
he wanted to provide the reader with an adequate description of the linguistic
context of France, especially since he intended to demonstrate how a traveller
may ‘arrive to the practicall knowledge of the languages’, as he put it in the title of
his travel guide.
In short, early modern scholars distinguished between different kinds of dia-
lects, thus expanding their conceptual apparatus in order to cope with the
complex and multi-layered phenomenon of linguistic variation. Not only classi-
fication was a main goal of this conceptual development, however. It often also
served to propagate the selected variety of a vernacular language that was being
standardized as the correct linguistic form; the selected variety received a quality
label, that of ‘principal dialect’. The early modern insistence on the existence of
different kinds of dialects reflected linguistic realities, most importantly the fact
that linguistic distance comes in degrees as well as the ongoing standardization of
the Western European vernaculars.
8.4 Mutual intelligibility: an early modern criterion
So we therefore say that the Jerusalemite tongue was distinct from the
Galilean in the manner of dialect. The reason is evident, since
Galileans could be distinguished from Jerusalemites by their speech,
certainly not because the language was different (otherwise they
would not have understood each other clearly), but the dialect was.¹⁵
To paraphrase this seventeenth-century observation on Hebrew diversity:

speakers of related dialects can understand each other, whereas speakers of
¹⁵ Pfeiffer and Martini (1699: 212): ‘Sic ergo linguam Hierosolymitanam à Galilæa ad modum
dialecti distinctam fuisse dicimus. Ratio evidens est, cum ex sermone Galilæi ab Hierosolymitanis
dignosci potuerint. Non certe, quod lingua diversa esset, (aliàs se invicem planè non intellexissent) sed
dialectus’.
118  ’  
distinct languages cannot. In modern linguistics, the validity of this criterion,

called mutual intelligibility in short, has been often discussed and generally
considered doubtful at the very least, even though many modern scholars
still rely on it (see Chambers and Trudgill 1998: 3–4; Chapter 20). How did
early modern scholars approach mutual intelligibility, or lack thereof, in
relation to the conceptual pair? Sixteenth-century authors generally did not
employ it as a criterion to determine whether a speech form was a language or
a dialect. Mutual (un)intelligibility was nonetheless eagerly used to character-
ize the relationship between specific language varieties. What is more, it is
striking how often the absence of mutual understanding among speakers
of related dialects was emphasized. As I have mentioned in Chapter 7
(Section 7.1), Martin Luther already observed in the first half of the sixteenth
century how speakers of German, living only thirty miles apart, were not able
to understand each other. The renowned Italo-French classical scholar Joseph
Justus Scaliger (1540–1609) suggested that mutual intelligibility was endan-
gered mainly when two lower-class people speaking different dialects were
forced to converse, for instance a peasant from Liège and an artisan from
Paris. Even though they both spoke varieties of French, they did not under-
stand each other and even seem to have spoken entirely different languages
(Scaliger 1610b: 123–4).
In other cases, different varieties of a specific language were described as mutually
comprehensible. The Greek scholar Theodosius Zygomalas (1544–1607), for
instance, explained that the existence of different dialects in vernacular Greek did
not prevent mutual understanding, seemingly suggesting that his observation was a
more general principle (Zygomalas in Crusius 1584: 99). A sixteenth-century
German grammarian appears to have employed mutual intelligibility as a criterion
to classify the dialects of his native tongue into two different groups, ‘High German’
(‘Oberländisch’) and ‘Low German’ (‘Niderle[n]disch teutsch’); he clarified that the
dialects that were part of the former group were ‘mutually intelligible’ or ‘inuice[m]
intelligibiles’ in the Latin original (Albertus 1573: .8).
The interpretation of the conceptual pair in terms of mutual intelligibility often
went hand in hand with other understandings of it, not in the least the Aristotelian
criterion. What is more, it was sometimes regarded as a logical consequence of it.
Indeed, if two varieties differed in terms of substance, they were not mutually
intelligible and, by consequence, distinct languages. If two varieties exhibited only
variation in terms of their accidents, they were mutually intelligible and, by
consequence, related dialects. Both conceptions appeared in close connection in
Georg Stiernhielm’s ideas on the language/dialect distinction, which are analysed
in a separate chapter because of their richness and complexity, and because they
mark a transition to a subsequent phase in the history of the conceptual pair (see
Chapter 12).
.    . . . 119
8.5 Johannes Goropius Becanus and immediate

mutual intelligibility
Even though mutual intelligibility was keenly used in describing the relationship
between specific dialects in the sixteenth century, it did not figure in more general
discussions of linguistic diversity, with one major exception: the work of the
Brabant physician Johannes Goropius Becanus (1519–73; Figure 8.2), infamous
for his theory that Dutch was the primeval tongue of Adam and Eve, put forward
in an age where most scholars attributed this privilege to the Hebrew tongue.¹⁶
Mutual intelligibility played a role of primary importance in Goropius’s distinc-
tion between ‘different languages’ and ‘identical languages’:
Now, since it has been shown what dialect means to Aristotle and what it means
to us, it should be enquired which dialects or languages are to be considered
Figure 8.2 Johannes Goropius Becanus by Philips Galle, 1572

Source: Rijksmuseum, Amsterdam. Public domain
¹⁶ See Van Hal (2010a: 77–139) and Frederickx and Van Hal (2015: 111–71) for nuanced discus-
sions of Goropius’s linguistic ideas.
120  ’  
identical and which different. The Greeks—in their fashion of speaking—

subsumed five different dialects under a single Greek language. We—in the
popular fashion of speaking we follow here—believe that it should be determined
thus: we call different those languages that mutually differ in such a manner that
he who understands one language does not immediately understand the other
language, too. However, we call identical those languages that, even though they
differ somewhat, nevertheless do not deviate from each other to such an extent
that their difference makes conversation and communication impossible.¹⁷
Clearly, Goropius regarded the Latin terms lingua and dialectus as synonyms,
choosing to base his metalinguistic distinction on such qualifications as ‘different’
and ‘identical’. Indeed, Goropius concentrated on determining whether specific
varieties should be understood as constituting either one and the same language or
different languages. His main criterion for doing so was immediate mutual
intelligibility, immediate in the sense that speakers had no time for reflection or
prior knowledge. In what followed, Goropius exemplified his criterion, arguing
that on this ground the Brabantian tongue was distinct from Swabian, just as
Spanish was from French and Italian. Brabantian and Flemish, on the other hand,
were identical tongues. Even though they were somewhat different, speakers of
these varieties could communicate with each other without problems.
It should be noted that Goropius seems to have attributed a language/dialect
distinction to the Greeks, who, he supposed, subsumed several dialects (dialecti)
under their language (sermo). He reformulated the allegedly Greek conceptual
pair in ‘the popular fashion of speaking’ as a difference between ‘different’ and
‘identical languages’. Goropius thus adopted a slightly different perspective, but in
essence his distinction can be said to boil down to a distinction between language
and dialect. After all, he suggested himself that it was merely a matter of termin-
ology. What is more, elsewhere in his work, Goropius appears to have employed
the term dialectus in the sense of ‘variety subsumed under a language, immediately
intelligible for speakers of related varieties’ in alleged agreement with Greek
terminology. For instance, he frequently used formulations such as ‘different
dialects of the same language’.¹⁸ Such a usage is in direct contradiction to the
fact that Goropius had initially defined the word dialectus as ‘the faculty of
¹⁷ Goropius Becanus (1580: Hermathena 3–4): ‘Iam quoniam quid Dialectus Aristoteli, quid nobis
sit, ostensum est; quærendum, quæ dialecti siue linguæ eædem, quæ diuersæ sint habendæ. Græci,
more suo locuti, quinque diuersas dialectos sub vno sermone Græco complexi sunt: Nos populari
loquendi more, quem hîc sequimur, ita statuendum arbitramur; vt eas linguas diuersas dicamus, quæ
sic inter se discrepant, vt non statim qui vnam intelligit, alteram quoque intelligat. Easdem verò; quæ
licet nonnihil differant, tantum tamen non recedit altera ab altera, vt discrimen earum colloquendi
communicandiq́ue auferat facultatem’.
¹⁸ Goropius Becanus (1580: Hispanica 5): ‘Alibi commemoraui nos Man dicere, Allemanos in
eadem significatione Mon, & id genus variationes esse in diuersis eiusdem linguæ dialectis quàm
frequentissimas’.
.    . . . 121
speaking in a native way about everyday matters’, putting it on a par with lingua.¹⁹
This friction can be explained by the fact that dialectus had a double sense for
Goropius, referring, on the one hand, to a specific historical variety of a language
and, on the other, to human language as a means of expression. Goropius’s case
thus constitutes an excellent example of the way in which the polysemy of
dialectus could impact on a scholar’s linguistic thought and his terminological
flexibility.
Goropius’s criterion of immediate mutual intelligibility enjoyed some success in
the seventeenth century and was adopted by several scholars who never men-
tioned Goropius’s name in this context, probably because of the controversial
nature of his views on Dutch linguistic primacy. The most striking case is that of
the Protestant theologian Johann Caspar Myricaeus (†1653), who in his Syriac
grammar copied Goropius’s passage on the distinction between different and
identical languages verbatim, but did not refer anywhere to Goropius’s work.²⁰
However, not everyone agreed with Goropius that related dialects were directly
mutually intelligible. Abraham Mylius (1612: 85), for instance, who was well
acquainted with Goropius’s work, remarked that a speaker of Syriac needed
training in order to immediately understand Arabic or Hebrew, this in spite of
the fact that these tongues differed only in terms of dialect. The criterion of
immediate mutual intelligibility appeared with a remarkably low frequency in
eighteenth-century thought. An important exception was the Swedish educator
Sven Hof ’s (1703–86) Dialectus Vestrogothica, a monograph on the Västergötland
dialect of Swedish.²¹ In it, Hof (1772: 11) claimed that a lack of mutual intelligi-
bility ‘without any peculiar instruction’ (‘sine peculiari institutione’) indicated
different languages, whereas related dialects were mutually understandable. More
generally, the interpretation in terms of mutual intelligibility was relatively rare in
eighteenth-century discussions of the language/dialect pair, which may be con-
nected to the increasing focus on specific linguistic features in analyses of the
conceptual distinction (see Chapter 15).
The emphasis on immediate mutual intelligibility by Goropius and others
suggests that a number of early modern scholars did not take it to be a binary
feature but rather saw it as a property with at least three possibilities: speech forms
can be either mutually unintelligible, or mutually intelligible after instruction or a
period of contact, or immediately mutually intelligible. This insight did not,
however, lead scholars to problematize the fact that neighbouring dialects are
¹⁹ I have paraphrased the definition from Goropius Becanus (1580: Hermathena 3): ‘ . . . facultas, vel
vsu, vel præceptis, vel ambobus acquisita; qua qui præditus est, de quauis re, in populari consuetudine
tractari solita, apud nationis cuiuspiam homines emendatè loqui possit’.
²⁰ Myricaeus (1619: }.2–3): ‘Porro lingvas diversas eas esse dicimus, quæ sic inter se discrepant, ut
non statim qui unam intelligit, & alteram possit: easdem quæ, licet non nihil differant, tantum tamen
una ab altera non recedit, ut discrimen earum colloquendi communicandique auferat facultatem’.
²¹ On Hof ’s place in the history of linguistics see Droixhe (1978: 123–4); Hovdhaugen et al.
(2000: 93–4).
122  ’  
more likely to be mutually comprehensible than related dialects that are geo-
graphically farther removed from each other. What is more, most authors seem-
ingly chose to ignore the fact that mutual intelligibility comes in degrees and took
the property in a strictly binary manner. The English bishop Brian Walton
(1600–61), for instance, did so in his prolegomena to the polyglot Bible he edited
in 1657, which touched on, among other things, issues of language history.
Speakers of related dialects could understand each other, but speakers of different
languages could not. That was at least what Walton guessed, as he felt compelled
to insert the Latin adverb forte, ‘perhaps’, into this claim of his. Still, he explicitly
stated that if there was no mutual intelligibility among speakers of two different
tongues, these would be improperly called dialects (Walton 1657: 3). Other
scholars seem to have been more confident about the binary nature of the
criterion. For instance, in his posthumously published remarks on the linguistic
diversity of the Holy Roman Empire, the French magistrate Guillaume Ribier
(1578–1663) maintained that speakers of different languages could not under-
stand each other, neither in speaking nor in writing (1666: 27).
The awareness of early modern authors that dialectal variation comes in
degrees, implicit in the insistence on immediate mutual intelligibility, manifested
itself more strongly in other ways, most notably in the assumption that there were
different kinds and levels of dialects, as I have argued earlier in this chapter.
8.6 The communicative reach of dialects versus languages
The emphasis certain authors put on the lack of mutual intelligibility among
related dialects suggests that these linguistic varieties were believed by some
scholars to have a restricted communicative reach. Conversely, a wide communi-
cative reach was from time to time posited for a language, as, for instance, by an
English bishop in a seventeenth-century discourse on the celebration of Mass in
an unknown tongue:
Where there are these different Dialects, there generally is one way of speaking,
which either from the eloquence, or fashionableness of it, so far prevails, as to be
the Standard of the Tongue, and to be used in Writing Books, Letters, &c. and is
understood by all. (Williams 1685: 5)
The extensive communicative reach of a language was connected to its written

status, causing it to be more permanent than the ever fluctuating dialects, most of
which were not used in writing.
In the eighteenth century, the emphasis on the extensive communicative reach
of a language, often specified as the common language, intensified. The Greek
émigré Alexander Helladius (1686–after 28 August 1714), for instance, stated the
.  123
following on the vernacular Greek and German languages in his 1714 description
of the Greek Orthodox Church:
Thus, in the modern era, there is a single common ecclesiastical dialect across the
whole of Greece, intelligible by all and at the same time not easily imitable by any
plebeian. This is not different from that Attic throughout Germany, which is
called “the High German par excellence”.²²
Helladius, although using the term dialectus both for the common variety and for
the vernacular dialects, contrasted the communicative reach of the common
variety to the miscommunication the dialects of the populace could cause. The
latter theme was treated at length in Helladius’s book by means of a series of
amusing anecdotes drawn from his own experience in Greece and Germany
(Helladius 1714: 192; Van Rooy forthcoming).
8.7 Conclusion
A key interpretation of the conceptual pair that emerged during the first part of
the early modern period was what I have been calling the ‘Aristotelian criterion’.
The idea that languages differed from each other substantially, whereas related
dialects varied only accidentally, was actively used as a diagnostic criterion to
determine the status of a speech form. It was also used for other, highly specific
purposes, for instance to denigrate someone’s polyglot competence. The
Aristotelian interpretation featured seldom in generalizing discussions, perhaps
because it was supposed to be self-evident, much like the language/dialect distinc-
tion itself. Its relationship to the Greek tradition was moreover ambiguous. Even
though the criterion was undoubtedly an early modern innovation, it was obvi-
ously inspired by traditional philosophical categories associated with Aristotle.
The friction between the binary nature of the language/dialect pair and the
realization that dialectal variation was a gradational property remained unre-
solved. Scholars posited different hierarchical levels within dialectal diversity but
refrained from elaborating on the question of how this multi-layered analysis of
linguistic diversity fitted in with the conceptual pair in general and its Aristotelian
interpretation in particular.
Closely associated to the Aristotelian interpretation was mutual intelligibility.
The dialect or language status of a speech form depended on the question whether
or not speakers could understand each other. Some early modern scholars relied
²² Helladius (1714: 187): ‘Ita moderno tempore, per universam Græciam unica est dialectus
ecclesiastica communis, & ab omnibus intelligibilis, non idem facile à quolibet plebejo imitabilis, non
aliter quam per universam Germaniam Attica illa, quæ κατ’ ἐξοχὴν das Hochteutsche vocatur’.
124  ’  
on this feature as an absolute criterion. Yet there were several authors who,
dissatisfied with a binary interpretation of the feature, specified the criterion as
immediate mutual intelligibility. It is rather paradoxical that some authors argued
that dialects of one and the same language were mutually understandable, whereas
others emphasized the limited communicative reach of dialects and the lack of
mutual understanding among speakers of closely related dialects. The realization
that dialects were not always mutually intelligible was moreover in contradiction
to the widespread early modern idea that dialects varied only on an accidental
level. An author’s emphasis on mutual unintelligibility among dialects was often
related to contextual factors. For instance, Martin Luther’s stress on the lack of
mutual understanding among the speakers of German dialects must be viewed in
connection with the confessional goals his linguistic programme had to serve.
Luther wanted to justify his adoption of the common language of the Saxon
chancellery as the linguistic medium for his German Bible translation; pointing
out the great linguistic diversity in Germany and the communicational problems
that went with it was part of that justification.
9
A subjective touch
Language beats dialect
Early modern scholars at first limited themselves to formulating more or less

objective interpretations of the language/dialect pair, in the sense that they did not
associate either concept with normative and evaluative properties. In the seven-
teenth century, more subjective voices appeared on the scene, and many scholars
degraded dialect to a status inferior to that of language. How and why did this
evolution occur? In the present chapter, I set out to answer these questions by
treating two major interpretations of the language/dialect distinction centring
around notions of analogy/anomaly and superiority/inferiority.
9.1 Analogical norm or anomalous deviation?
The common language is that which follows the more exact analogy
or, as usually happens with extinct languages in particular, the norm
of grammatical precepts, from which the dialects variously diverge
now and then.¹
According to this Wittenberg disputation on the causes of dialects, a common

language reflected a regulated and analogical norm, whereas dialects were char-
acterized by anomalies deviating from that norm.² This interpretation of the
conceptual pair had its roots in the close of the sixteenth century. For instance,
the Protestant philologist Otto Walper (1543–1624), professor of Greek and
Hebrew in Marburg, made the idea of an analogical language and its anomalous
dialects explicit in his handbook for the Greek dialects. This work was first
published in 1589 as an independent treatise; a year later Walper incorporated
it into his Greek grammar, which enjoyed numerous reprints for several decades.
In the dedicatory letter prefixed to his handbook, Walper pointed out the
following:
¹ Schwartz and Helm (1702: .1): ‘communis lingua est, quae exactiorem analogiam, aut, quod in
intermortuis linguis maxime fieri solet, praeceptorum Grammaticorum normam sequitur, a quibus
dialecti uarie aliquando diuertuntur’.
² On analogy and anomaly in early modern linguistic thought and their ancient roots see Haßler
(2009a).
DOI: 10.1093/oso/9780198845713.001.0001
126   
Even though indeed almost all grammarians add the dialects as if they were some
anomalies after the analogy of the precepts for the common language, so that it
seems to be quite superfluous to prepare a separate treatment of the dialects,
still, if we would examine the matter more thoroughly, we will notice that there
will be a rather different conclusion in view of the condition and utility for the
studious youth.³
Walper apparently attributed the contrast between the anomalous Greek dialects
and the analogical Koine to the methods and grammatical descriptions of his
predecessors, in whose works such an opposition was, however, only implicit at
most. At the same time, he deplored and tried to remedy the idea that dialects
were arbitrary deviations from the regulated common language. All the same,
such an opposition was still rare in the sixteenth century. What is more,
anomaly was not at all considered to be inherent in the concept of dialect. The
English theologian Thomas Stapleton (1535–98), for example, was convinced that
the Greek dialects ‘might be gathered and collected in to some Orders of Rules’
(1566: 59).
The seventeenth century witnessed a profound change in conceptualizing the
language/dialect distinction in highly normative terms. The influential German
grammarian Justus Georg Schottel revealingly observed that ‘in all dialects, there
is something faulty, which cannot occupy the position of rule in the language
itself ’.⁴ Elsewhere, too, Schottel (1663: 158) emphasized the impossibility of
formulating rules for dialects, which he conceived as intrinsically variable. In
other words, language was the regulated norm or point of reference from which
its dialects anomalously deviated. Even though Schottel did not mention the
Greek dialects at all in this context, this normative interpretation seems to have
originated in Byzantine grammatical practice. Certain Byzantine scholars had
opposed the Koine to the dialects at various occasions, suggesting at times that
the Koine was analogically constructed, whereas the dialects were anomalous
variations on the Koine. They did so, however, without working with an abstract
language/dialect-like distinction. This approach was silently adopted by
Renaissance Hellenists to become explicit only towards the end of the sixteenth
century in the work of Walper, who criticized the grammarians’ default attitude.
In the course of the seventeenth century, the analogy/anomaly opposition was
generalized and detached from the context of Greek studies. As a result, it became
applicable to a wider range of languages and absorbed into the conceptual pair.
³ Walper (1589: †.6): ‘Etsi verò ab omnibus penè Grammaticis post analogiam præceptorum
linguæ communis, Dialecti tanquam anomaliæ quædam subijciantur [sic], ita vt fermè superuacaneum
esse videatur seorsim de Dialectis tractationem instituere: tamen si rem penitius introspiciamus,
diuersum potius ex re & vtilitate studiosæ iuuentutis fore animaduertemus’.
⁴ Schottel (1663: 174): ‘Omnibus dialectis aliquid vitiosi inest, quod locum regulæ in Lingua ipsa
habere nequit’.
.      127
Of course, language regulation was an evident and central theme for vernacular
grammarians even before the seventeenth century, but it was only after 1600 that
more sternly normative attitudes came to dominate language disputes. In France,
in particular, scholars intensely debated ‘good usage’ (bon usage) in the tracks of
François de Malherbe (1555–1628), a tradition in which the genre of the ‘remarks’
came to play a central role (Ayres-Bennett and Seijido 2011). Metalinguistic
considerations on normative concepts, including language and dialect, remained
relatively rare throughout the early modern period, however (cf. Haßler 2009b:
674). It seems nonetheless justified to conclude that early modern scholarship on
the Greek language stimulated to some extent more general observations on the
prescriptivist idea that dialects were anomalous deviations from the linguistic
norm. The strongly normative focus of early modern vernacular grammarians,
by the way, indicates that their main concern normally was with the relationship
between dialects and the language from which they were believed to deviate rather
than with the interrelationships of related dialects (cf. Haas 1994: ).
9.2 From common to standard language
Language was often further specified as common language when an early modern
scholar interpreted the conceptual pair in terms of analogy and anomaly. Out of
the idea that a common language constituted the regulated norm from which
dialects deviated, the concept of standard language seems to have emerged in the
second half of the seventeenth century, with a particular position for England.
The English bishop John Williams (?1636–1709) was one of the first to employ
the term standard in a linguistic sense, while mentioning at the same time some of
the basic features of what most modern linguists would call a standard language.
Williams did so in the context of a fundamental problem for the celebration of
Mass: in what language should it be read? He observed in this regard:
Where there are these different Dialects, there generally is one way of speaking,
which either from the eloquence, or fashionableness of it, so far prevails, as to be
the Standard of the Tongue, and to be used in Writing Books, Letters, &c. and is
understood by all. Such I conceive was anciently that which is called the common
Dialect in Greek: And of the like kind is that which is spoken in and about the
Court, and by Scholars and persons of a liberal education amongst us; and
elsewhere. (Williams 1685: 5; cf. Chapter 8, Section 8.6)
Which is (to speak charitably) for want of observing, that the Dialects are but
several modes of speaking the same Tongue; and that ordinarily there is some
common Standard, which (as I have said) over-rules the rest, and is a guide
common to all: As here in England, notwithstanding there be several Dialects,
128   
and that there is one in Scotland differs much from them all; yet there is but one
Translation of the Bible, and one Service for the use of the whole, and that is fully
if not equally understood by all. (Williams 1685: 8)
In Williams’s opinion, the standard variety was selected because of its ‘eloquence’
or ‘fashionableness’, as it was the only one appropriate for written communication
and it had the advantage of being intelligible to all speakers of the language. To
clarify his standard language concept, Williams referred to both the Greek Koine
and the high variety of English, spoken at the English court and by the intelli-
gentsia. The standard variety not only ‘roofed’ the other dialects, to use modern
linguistic terminology, but was also an example for them—or in Williams’s
phrasing: it overruled and guided them. In other words, the dialects were socially,
culturally, and linguistically subordinate to the standard, which was employed by
the upper classes and which was the ideal the dialects should aim to attain.
Normative thinking intensified in the course of the eighteenth century. For
instance, an English classical scholar maintained that ‘the distinction of dialects
can be only known to a cultivated and, in some degree, settled state of language, as
deviations from an acknowledged standard’ (Wood 1775: 238). In the French
tradition, a number of scholars distinguished between two kinds of varieties
subsumed under a language; they were either legitimate and regulated dialects,
occurring in politically diversified areas, or corrupted patois that lacked rules and
deviated from the normative variety of the centralized state, the latter obviously
applying to the kingdom of France (cf. Chapter 16, Section 16.2). Their concep-
tion and evaluation of a regional variety of a language, in other words, depended
on the political context.
Language was the analogical norm from which dialects were anomalous devi-
ations. Yet this opposition was not set in stone, as some early modern scholars
were aware that certain speech forms which they labelled languages were in fact
not much more than what one might call ‘upgraded’ dialects. They assumed, in
other words, that language status was something that could be achieved by the
standardization of a dialect. Modern research has pointed out that this process
consists in the selection of a specific variety as the norm, its grammatical codifi-
cation, its elaboration, and its wide acceptance as the standard (Chapter 4,
Section 4.1). It is, however, possible to find traces of awareness of some of these
standardization phenomena in various early modern sources. For instance,
scholars noticed that most languages had a principal dialect, the dialect selected
as the linguistic norm, as, for instance, the Spanish scholar Juan Luis Vives did
already in 1533. Vives (1533: .iii) mentioned Attic for Greek, Castilian for
Spanish, Tuscan for Italian, and Parisian for French, all of which he presented
as regional varieties of their respective tongues enjoying a privileged status.
Codification and elaboration were widely considered indispensable for turning
a vernacular dialect into a full-fledged cultivated language that could compete with
.    ? 129
Latin and other major vernaculars (cf. Burke 2004: 89). The German poet and
lexicographer Nathan Chytraeus (1543–98), for example, sounded rather desperate
about the linguistic state of affairs in Germany when he wrote the following in the
dedicatory letter prefixed to his pioneering Latin–Low Saxon lexicon of 1582:
To be sure, no matter what dialect of no matter what language has certain

elegancies and peculiar ornaments of its own. Doubtlessly these need to be
carefully studied and collected by someone who, in the German or whatever
other nation, would desire to establish a common language, such as there once
was in Greece, and to adorn it with characteristic and clear words and manners of
speaking, and besides also with figures, as if it were with jewels and flowers.⁵
The concept of common language, understood as a prestigious means of commu-

nication among the elites of a nation, played a key role here, since it was viewed
as the goal to achieve. Indeed, Chytraeus expressed the hope that German would
make up lost ground in this regard by following the example of the ancient
Greeks and other nations of his time. In Chytraeus’s view, however, the com-
mon language to be designed did not consist in one single dialect but rather in a
compilation of the best features of different dialects, which reveals a distinctly
positive attitude towards regional variation. Yet dialect was soon to acquire an
overly depreciatory interpretation.
9.3 Superior or inferior?
 . . . Particular speech of a province, corrupted from the gen-

eral or principal language of the kingdom.⁶
The lemma dialecte in Antoine Furetière’s (1619–88) influential French dictionary

suggests that, to the lexicographer’s mind, language had positive properties,
whereas dialect was primarily linked with negative features. A language was
considered superior, pure, and adequate for literature, but a dialect was inferior,
corrupted, and not at all fit for literary purposes. This interpretation was closely
tied up with the opposition analogy/anomaly and followed almost naturally from
it. Language was analogical and therefore superior to a dialect, which was
⁵ Chytraeus (1582: .3-): ‘Habet enim quælibet etiam cuiuslibet linguæ dialectus, suas quasdam
concinnitates, sua peculiaria ornamenta, quæ sanè accuratè consideranda & colligenda forent ei, qui in
natione Germanica, vel alia quacunq[ue], communem aliquam, qualis in Græcia fuit, linguam con-
stituere; eamq́[ue] vocabulis & modis loquendi proprijs & perspicuis; & præterea figuris quoq[ue]
tanquam gemmis & flosculis cuperet exornare’. On Chytraeus’s work as a lexicographer see Considine
(2017: 116–17).
⁶ Furetière (1690: s.v. dialecte): ‘DIALECTE . . . Langage particulier d’une Province, corrompu de la
Langue generale ou principale du Royaume’.
130   
anomalous and therefore of lower esteem than a language. This very subjective
interpretation of the conceptual pair was nearly absent from sixteenth-century
thought. The closest we come to it is in a work by the English literary critic George
Puttenham (1529–90). In The arte of English poesie, Puttenham (1589: 119–20)
opposed the inalterable language of a nation to speach, that was variable in terms
of cultivation, literary usage, and regulation (see Hickey 2010: 15). Still, most
sixteenth-century authors did not regard dialects as inherently inferior entities,
even though individual dialects often received various evaluative rubrics such as
‘elegant’, ‘coarse’, ‘rustic’, ‘smooth’, and so on.
A negative interpretation of the dialect pole of the conceptual opposition might
have initially been obstructed by several factors. Firstly, dialectus was still very
closely connected with the literary varieties of the revered Greek language in the
sixteenth century. Secondly, the highly valued selected variety of a language was
often still termed dialectus in the first half of the early modern period, precluding a
downgrade of the concept of dialect. Thirdly, local varieties were not yet auto-
matically dismissed as inferior entities. Indeed, several authors even stressed the
richness dialects could exhibit. The printer-scholar Henri Estienne (1582: *.iii-),
for instance, described how certain dialect features could contribute to the splen-
dour of the French language like beauty marks on a pretty face. Even more
tellingly, Estienne (1579: 133–4, 147) suggested that dialects related to a language
as a rich man’s country seats to his more luxurious main residence in the city (cf.
Swiggers 2009: 73). This comparison implied that even though dialects were not
inherently corrupt, they were nevertheless somehow inferior to the normative
language of a country, closely associated with the capital. Estienne did loosely
associate dialect with vice elsewhere in his linguistic work, for instance when he
criticized the French grammarian Jean Pillot (1515–92) for mistaking a fault in his
dialect for a grammatical rule (Estienne 1582: 202; see Bruña Cuevas 1998: 528).
He moreover emphasized that one could only speak of a dialect’s richness insofar
as the depravations and corruptions of ‘the insignificant populace’ (le menu
peuple) had been cleared away. For instance, Parisian speech served as the basis
for the ‘pure and native French language’ (pur et nayf langage françois), but this
elegant variety, too, had to be cleansed from vulgar elements (Estienne 1579:
135, 143).
More in general, sixteenth-century scholars did not consider a dialect to be
intrinsically typical of lower social classes. For in their time, standardization
processes were still under way, and most members of the higher classes still
spoke in their local dialects, even though they increasingly tried to write in the
common language that was being designed. This sociolinguistic situation explains
why humanists like Estienne attributed a prominent role to the dialects in
enriching the standard; local varieties benefitted from the social prestige of their
speakers and were therefore a factor to be reckoned with, but the fate of the
dialects was soon to change.
.    ? 131
During the seventeenth century, especially in its second half, corruption,

degeneration, and lack of cultivation were ever increasingly associated with
dialects, whereas language was conceived as a superior entity linked to purity.
An early example of this development can be found in The ambassador, a Latin
treatise on diplomacy first published in 1598 and authored by the Italo-French
diplomat Carlo Pascale (1547–1625). In this work, Pascale answered the question
of what language should be used by a diplomatic mission as follows:
As far as I am concerned, I think it should be in that language which is native to

the nation of the embassy . . . I am talking here about the embassies of those
nations that have their own language, as it is offending to call the inelegant,
worthless, and obscure idioms of some provinces languages. I prefer to call them
servants of languages.⁷
Pascale, well-versed in the Greek language and literature, referred to these

inferior linguistic entities as provincial ‘idioms’ (idiomata), using a term that
was not as closely connected with the prestigious Greek literary varieties as
dialectus was. This terminological choice made it possible for him to operate
with a conceptual opposition consisting of an inferior and a superior pole: idioma
versus lingua. Pascale was, in other words, actively looking for a way to distinguish
two levels of linguistic entities, a superior and an inferior one, suggesting that
there was already a need for it but not yet an established metalinguistic apparatus
with which to do so.
The idea that dialects were faulty varieties of a language, prominent from the
second half of the seventeenth century onwards, only made sense if the literary
Greek dialects no longer had a privileged connection with the concept of dialect.
One of the most telling testimonies to such a view is Antoine Furetière’s definition
of dialecte, explained as ‘particular speech of a province, corrupted from the
general or principal language of the kingdom’, as quoted above. Furetière inter-
preted the term patois even more negatively, defining it as ‘corrupted and churlish
speech, such as that of the insignificant populace, peasants, and children who are
not yet able to pronounce well’.⁸ Dialecte and patois were, in other words,
semantically closely related in Furetière’s eyes, even though patois was primarily
associated with social differences and dialecte was understood as a corrupted
regional variety of a language.
In the eighteenth century, the idea that dialects were corrupt entities led some
scholars to even promote the abolition of certain dialects. In an anonymous essay
⁷ Pascale (1598: 269–70): ‘Equidem puto illa quæ iis vernacula est, quorum legatio est . . . Hîc loquor
de legationibus earum gentium, quibus est propria lingua. Nam incompta, egena, et contorta idiomata
quarumdam prouinciarum tædet vocare linguas. Malo linguarum pedissequas’.
⁸ Furetière (1690: s.v. patois): ‘Langage corrompu & grossier, tel que celuy du menu peuple, des
paysans, & des enfants qui ne sçavent pas encore bien prononcer’.
132   
of 1743, for instance, it was argued that Low Saxon should be wiped out in favour
of High German and for the greater glory of the empire (Anon. 1743). Others,
however, restored the link with the Greek background to a certain extent and
came to regard dialects as full-fledged and legitimate varieties of a language,
distinguishing them from the corrupted varieties of a language, termed jargon
or patois (Chapter 16, Section 16.2). A number of scholars projected the idea of
corruption and degeneration on the Ancient Greek dialects, too, as did John
Wilkins (1614–72), who wrote at the very beginning of his well-known Essay
towards a real character, and a philosophical language of 1668:
The Greek was anciently of very great extent, not onely in Europe, but in Asia too,
and Afric, where several Colonies of that Nation were planted; by which disper-
sion and mixture with other people it did degenerate into several Dialects.
(Wilkins 1668: 3)
One seventeenth-century Swiss Hellenist even suggested that the Greek dialects
were sheer incorrigible deviations of the Greek language (Wyss 1650: 2).
It should be noted that the opposition superiority/inferiority was not perceived
as a strictly binary one, as statements such as the following reveal: ‘the more
cultivated a vulgar form of speaking is, the more closely it approaches the
common language’.⁹ What is more, the cultivated language could become tainted
by the inferior dialects. According to the German polymath Daniel Georg Morhof
(1639–91), ‘in such a vast diversity of dialects, it can easily happen that the more
cultivated language receives some dirt’.¹⁰ The corruptive influence of dialects on
the high variety could, however, be remedied, Morhof argued, by political cen-
tralization, which would eventually result in linguistic normalization as well
(Morhof 1685: 146).
It is remarkable that one seventeenth-century French author, the French Jesuit
Pierre Besnier (1648–1705), observed that the most polished language was often
the most corrupted one, whereas the patois were pure (1674: 23–5). Besnier’s
opposition centred around the artificial nature of a cultivated language versus the
pristine and primitive quality of the patois. Here, we have a rare example of an
early modern scholar rightly suggesting that a standard language intrinsically
postdates a situation of dialectal diversity. Other authors formulated different,
less historically accurate solutions to the question as to how the standard language
related to its cognate dialects in terms of time. The grammarian Justus Georg
⁹ Schwartz and Helm (1702: .1): ‘Quo cultior autem aliqua dicendi forma uulgaris est, eo propius
ad . . . linguam communem accedit’.
¹⁰ Morhof (1685: 113): ‘Nos tamen inde observabimus, quam facile sit in tantâ diversitate dialec-
torum cultiorem linguam aliquid sordium adsciscere’.
.    ? 133
Schottel (1663: 174), for instance, suggested that scholars had restored the High
German language from the existing dialectal diversity to its original form.
The association of dialect with degeneration, corruption, and lack of cultivation
reflected a number of sociolinguistic evolutions occurring in seventeenth-century
Western Europe. Most notably, the high and middle classes—royalty, nobility,
intelligentsia, bourgeoisie—went out of their way to write and eventually speak the
normative variety, identified as the language. In Germany, for instance, ‘erudite
men and those imitating them principally use [the common language] in writing,
yes indeed also—as far as it can happen—in pronouncing’.¹¹ In other words,
speaking a dialect was no longer solely a geographical but increasingly also a
social mark. As a matter of fact, when confronted with the high variety, one’s
native dialect could even become a source of shame—or to put it with an early
modern English saying: ‘She hath been at London to call a strea a straw, and a waw
a wall’. ‘This’, the renowned English lexicographer John Ray (1627–1705)
explained, ‘the common people use in scorn of those who having been at
London are ashamed to speak their own Countrey dialect’ (1678: 75).
Dialects were ever more closely associated with the peasantry and were opposed
to the cultivated language of high society, as the testimony of an early seventeenth-
century English grammarian reveals:
What I say here regarding the dialects, you must realize, refers only to country
people, since among persons of genteel character and cultured upbringing, there
is but one universal speech, in pronunciation and meaning.¹²
Similarly, the German grammarian Schottel (1663: 174) reserved the actual
language for the men of high social classes, who played an active role in its
formation and elaboration. The insistence on this idea seems to have been
strongest in the French context, which is probably connected to its high degree
of political as well as linguistic centralization (e.g. Rousseau 1781 [1755]: 243–4).
In short, whereas dialect did not have an intrinsically negative connotation in
the Cinquecento, the concept suffered from a devaluation in the seventeenth and
eighteenth centuries. It occurred not only in countries with a high degree of
political centralization such as France, but also in politically diversified areas
such as the states of the Holy Roman Empire. The increasing frequency of the
interpretation can probably be viewed as a consequence of the growing emanci-
pation of the dialect concept from the Greek heritage with its valued literary
¹¹ Schwartz and Helm (1702: .1): ‘lingua communis Germanica . . . quam eruditi, et, qui hos
imitantur, potissimum in scriptione, imo et, quantum fieri potest, in pronunciatione, usurpant’.
¹² Gill (1619: 18): ‘Et quod hîc de dialectis loquor, ad rusticos tantùm pertinere velim intelligas: nam
mitioribus ingenijs et cultiùs enutritis, unus est ubique sermo & sono, & significatu’. The English
translation is adapted from Gill (1972: .104). For Gill’s views on the English dialects see Kökeritz
(1938).
134   
Table 9.1 Transfer of social prejudices to the conceptual pair

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------▸
Social class Prejudice Speech Conceptual pair
Higher Cultivated Standard and superior Language

Lower Unrefined Regional and inferior Dialect
The arrow indicates the direction of transfer
dialects as well as of the great rise in esteem of the emergent standard languages.
Indeed, the evaluative interpretation of the conceptual pair appeared around the
time that the selected standard variety was gradually being detached from specific
regions. The norm was ‘delocalized’, to quote Peter Burke (2004: 110), as part of a
broader ‘civilizing process’, involving primarily the European elites:
the glamour of the court and its influence on provincial elites encouraged the
adoption of the new form of language, which became a sign that its users were
distinct from and superior to ordinary people. The reform of speech was part of a
wider change, the withdrawal from participation in many forms of popular
culture on the part of European elites.
The evaluative interpretation of the conceptual pair was no doubt the result of
transferring preconceived ideas about specific social classes to the language variety
they used and eventually to the conceptual pair, too, a process schematized very
roughly in Table 9.1.
9.4 Conclusion
The distinction between language and dialect acquired a highly subjective dimen-
sion from the end of the sixteenth century onwards. A dialect was increasingly
conceived as an anomalous deviation from the analogical norm of the language.
This interpretation appears to have been inspired partly by assumptions about
language-external, social realities and partly by standardization processes, in
which non-selected dialects were usually only marginally involved, if at all. The
fact that non-selected dialects did not undergo processes of grammatical codifi-
cation and regularization no doubt led early modern scholars to associate dialect
with deviation from the codified variety and a lack of analogy. This idea is likely to
have paved the way for the interpretation in terms of status and was closely
intertwined with it. A dialect was anomalous and therefore inferior, seems to
have been the inference, whereas a language was analogical and by consequence
superior. A subjective turn occurred early in the history of the conceptual pair,
.  135
and early modern authors soon made a habit out of speaking in depreciatory
terms about dialects. The link between regional speech and lower social class
intensified, which no doubt fostered negative conceptions of dialect.
Remarkably enough, early modern scholars generally did not put the interpret-
ation of the conceptual pair in terms of analogy and anomaly into a diachronic
perspective. In other words, they did not usually make explicit that a codified,
regulated language could come only after a situation of dialectal diversity without
an acknowledged standard. Diachrony did, however, occupy a central position
in the final major interpretation of the conceptual pair that arose in the early
modern period.
10
The conceptual pair and language history
Language generates dialects
When humanists created the language/dialect pair in the sixteenth century, they
initially conceptualized it uniquely from a synchronic perspective; a dialect
resorted under a language at a specific moment in time. Soon a historical inter-
pretation of the distinction between language and dialect was devised, as numer-
ous early modern scholars began to consider a dialect to be a speech form
posterior to and genealogically deriving from a language. In this case, the latter
was often metaphorically termed mother language, the former daughter dialect.
When and why did this innovation with far-reaching consequences occur?
10.1 Language mothers and offshoot dialects
We can call language mothers those from which many dialects—

offshoots as it were—are derived. Offshoots of one language mother
are indeed connected by a certain correspondence, but there is no
mutual kinship among the language mothers, either in words or in
analogy.¹
With these observations the great classical philologist and professor at Leiden
university Joseph Justus Scaliger (Figure 10.2) opened his very short Diatribe on
the languages of the Europeans (Figure 10.1), originally a letter written to his
colleague Paulus Merula in 1599 and first published by Merula (1605: 271–2),
albeit against Scaliger’s will. Five years later, the letter was included in a posthu-
mous collection of Scaliger’s unpublished writings. In his Diatribe, Scaliger con-
troversially argued that there was no kinship whatsoever among the eleven
language mothers of Europe, a remarkable idea that soon met with fierce criticism
(see Van Hal 2010b). What is of interest to us, however, is what his opening
statement reveals about the metalinguistic concepts with which he was operating.
To Scaliger’s mind, dialects derived genealogically from an originally unitary
¹ Scaliger (1610a: 119): ‘L Matrices vocare possumus, ex quibus multæ dialecti, tanquam
propagines deductæ sunt. Propagines quidem vnius matricis linguæ commercio inter se aliquo
coniunctæ sunt: Matricum vero inter se nulla cognatio est, neque in verbis, neque in analogia’.
DOI: 10.1093/oso/9780198845713.001.0001
.      137
Figure 10.1 Joseph Justus Scaliger’s influential Diatribe on the languages of the
Europeans
Source: Archive.org. Public domain
language and were therefore secondary to it in terms of time. For this idea, Scaliger
was most likely inspired by the work of his predecessors, even if it is impossible to
name any precise sources. However, his usage of the Latin term propago, ‘offshoot,
descendant’, in close connection with dialectus may have been inspired by Conrad
Gessner’s observations on the Germanic family of languages. Scaliger might also
have been influenced by certain sixteenth-century texts in which the interpret-
ation was only latently present (see Chapter 4, Section 4.4; Chapter 5, Section 5.6).
It was, however, only in Scaliger’s wake that the genealogical interpretation of
the conceptual pair came into the limelight (cf. also Metcalf 2013: 52; Formigari
2004: 89). His oft-quoted Diatribe indeed seems to have been of central import-
ance for this evolution, and Scaliger may rightly be regarded as playing a pivotal
role in the emergence of the new language-historical conception of the language/
dialect distinction, still in vogue today. For instance, in his immediate environ-
ment, his colleague Paulus Merula assumed in 1605 that dialects were inherently
posterior to a language, as, for instance, when he referred to the problem of how
many original languages there were in the world:
138      
Figure 10.2 Scaliger c. 1607, engraving by Johann Theodor de Bry (1561–1623)

Source: Rijksmuseum through Europeana. Public domain
Let us not elaborate on the number (which I believe to have been lower than
Ephorus and others declare). This is certain, that those original languages, after
Hebrew, have produced many dialects by means of migrations hither and
thither.²
Here, Merula used the term dialectus for varieties genealogically related but
geographically dispersed.
Scaliger’s influence reached far beyond his intellectual network in Leiden,
however. The Anglo-Welsh author and traveller James Howell, for example,
likewise conceptualized dialect as a speech form historically derived from a
language. Howell was without doubt inspired by Scaliger’s ideas. In fact, he
repeatedly referred to Scaliger’s linguistic classification in his popular collection
of largely fictional letters published under the title Epistolae Ho-elianae in 1650.
² Merula (1605: 207–8): ‘De Numero (quem minorem, quam Euphorus ceterique edunt, fuisse
censeo) non laboremus: hoc certum, primigenias illas post Hebraicam multas Dialectos, Migrationum
huc illuc ope, peperisse’.
.      139
In an attempt at fine-tuning Scaliger’s account, Howell (1650: 78) at one instance

claimed there were fourteen ‘vernacular and independent tongues of Europe’ as
opposed to Scaliger’s eleven. Elsewhere, he suggested that the ‘mother toungs of
Europe are thirteen, though Scaliger would have but eleven’ (Howell 1650: 91).
Although disagreeing with Scaliger—and himself, for that matter—on the precise
number of ‘mother toungs’, Howell did adopt quite naturally Scaliger’s historical
conception of dialect throughout his writings. In his Instructions for forreine
travell of 1642, he argued, for example, that a good knowledge of Latin was
‘behooffull’, adding that it greatly facilitated the study of Italian, Spanish, and
French, which ‘are but Dialects or Daughters’ of it (Howell 1642: 21–2). Dialect
and daughter were conceived as metalinguistic synonyms, an equivalence that had
been largely absent from sixteenth-century usage.
An increasing interest in linguistic diversity, classification, and history led early
modern scholars to sedulously adopt the language/dialect pair in this new
language-historical sense, especially from the early seventeenth century onwards,
this in order to map out the historical interrelationships of the countless tongues
appearing on their horizon. It comes as no surprise that in such cases dialectus was
often accompanied by words meaning ‘offshoot’, ‘descendant’, and ‘daughter’.
A typical early modern statement in which the family metaphor appeared was,
for instance, the following: ‘the Latin language mother has produced as daugh-
ters and dialects Italian, Spanish, French, Rhaetian, Sardinian, and Romanian’.³
The use of botanical and family metaphors quickly gained momentum in
language studies, culminating in nineteenth-century historical-comparative lin-
guistics by conceptual cross-fertilization with other sciences, especially biology
(cf. Chapter 17, Section 17.3).
The view that a language generated dialects apparently contradicted the emer-
gent idea that dialects tended to exhibit archaic features. A prominent proponent
of the latter idea was Pierre Besnier, a French Jesuit whom I have introduced in the
previous chapter. In his interesting booklet The reunion of languages, or The art of
learning them all by means of only one, first published in Paris in 1674, Besnier
treated, among other things, the method of comparing tongues. In doing so, he
made the very uncommon suggestion that the cultivated varieties of French,
Italian, and Spanish were often the most corrupted forms of these languages,
and that among other things dialectal data could shed light on the obscured origin
of many of their words (Besnier 1674: 23–5). This view of Besnier’s raises an
important question: how can the view that dialects were archaic be reconciled with
the historical conception of dialect as inherently posterior to language? At first
sight, this idea seems to be a paradox, but it is only an apparent one, as language
was conceptualized by Besnier from a more or less synchronic perspective as the
³ Kirchmaier & Jäger (1686: 17): ‘Latina lingva mater filias & dialectos peperit Italicam, Hispanicam,
Gallicam, Rhæticam, Sardicam, VValachicam’.
140      
recently designed normative variety from which dialects deviated rather than as
the common ancestor of all the dialects. In fact, during the seventeenth century
scholars started to increasingly differentiate between these two conceptualizations
of language (cf. also Section 10.2), even though this distinction was usually not
reflected in their definitions.
The idea that a language diversified over time into related dialects naturally
invited scholars to frame this process within larger timespans. Key to this zooming
out was the assumption that related dialects gradually became so different from
one another that they should be viewed as distinct languages rather than as mere
dialects. The first scholar to problematize language change and diversity at length,
the Dutch humanist Abraham Mylius, made the following observation already
in 1612:
In fact, the more distance is left from that time and place when and where this
first change occurred, the greater also is the change that prevailed. This is why it
is so necessary that languages, because of the excessive distance of both elements
[sc. time and place], eventually degenerate entirely from different dialects into
other languages, which do not retain anything of the former language anymore
and barely even the likeness of a word.⁴
Mylius emphasized the vast differences existing between languages that were
originally dialects of one and the same language, even up to the point that there
were barely any similarities left. This observation did not prevent Mylius from
calling these different languages dialects. In other words, speech forms differing on
the level of language in Mylius’s time could be termed dialects because of their past
status as dialects of one language, even though they now exhibited considerable
differences. It is not difficult to see how this metonymic usage of the term dialectus
must have often led to a confusion of synchronic interpretations of the word with
its new language-historical meaning.
Remarkable about Mylius is that he did not refer to Scaliger’s Diatribe, even
though he entertained contacts with Leiden philologists and was acquainted
with some of Merula’s work (Van Hal 2010a: 210, 231). The early presence of
the language-historical interpretation of the term dialectus in Mylius’s Belgian
language can therefore be taken as indirect evidence in favour of his familiarity
with Scaliger’s work. There is, however, another possible explanation. Perhaps
Mylius was influenced by the works of Theodore Bibliander and Conrad Gessner,
which he had read (see e.g. Mylius 1612: 213). As I have contended in Chapter 5,
⁴ Mylius (1612: 85–6): ‘Quanto vero plus spatij ab illo tempore, quando, & ab illo loco, ubi prima illa
mutatio contigit, est interpositum, tanto quoque mutatio major obtinuit. Id quod ita usus est, ut linguæ
rei utriusque nimiâ distantiâ, ex diversis dialectis, tandem prorsus degenerent in alias, prioris nihildum
retinentes, & vix quidem vocis alicujus simulachrum’.
.    . . . 141
Section 5.6, it is easy to read a genealogical interpretation of dialectus into some

passages of these authors’ works, even though such a conception does not neces-
sarily chime with the authors’ intentions. Whatever the case, there was a large group
of late sixteenth- and early seventeenth-century authors who tried to develop a
workable concept to designate a variety, dialect, genealogically deriving from
another variety, language, a process that seems to have been strongly catalysed by
Scaliger’s widely read work and, more broadly, by the emerging interest in linguistic
diversity as a time-bound phenomenon.
10.2 An early critic of the language-historical interpretation
The language-historical interpretation of the conceptual pair soon faced criticism,

however, because it entailed the adoption of a different time perspective. The
other conceptions of the language/dialect distinction were all synchronic, in that
they concerned a snapshot in time, whereas the new meaning was the result of
putting language and dialect into a historical scheme. An early critic of this
language-historical usage of the conceptual pair was the Swiss orientalist and
theologian Johann Heinrich Hottinger (1620–67; Figure 10.3).
Having studied in his native land as well as in the Low Countries, including
Leiden, Hottinger occupied the chairs of theology and Oriental philology in
Zürich and Heidelberg. In 1667, he accepted a professorship at Leiden university,
but his boat capsized on his way there. In this accident, he drowned together with
three of his children. Some years before this tragic event, in 1661, he had published
one of his great works, his Oriental etymological dictionary, in which seven
languages featured. To this work, published in Frankfurt, Hottinger prefixed a
dissertation explaining the goals, usage, and methodology of his lexicon, in which
he also touched upon the relationships among the Oriental tongues included in
his work. His view on the position of the Lingua Sancta, the ‘holy language’, in this
linguistic group deserves to be quoted in full here:
After all, from [the comparative study of Oriental languages], it manifests itself
very clearly . . . that the holy language [Hebrew] is the mother of all other Oriental
languages and that the remaining are daughters, either more or less degenerate.
There are some who find pleasure in the nomenclature of dialect here. To
Christian Ravis, in his Disputatio de Dudaim, p. 29, “the Hebrew, Chaldean,
Syriac, and Arabic dialects appear to be one single language” [= Ravis 1656: 29].
“These dialects, which by the mistake of some people are now unfortunately
dispersed, are believed to differ by a huge amount, even though they are
connected like the Aeolic, Doric, Attic, and Ionic, the principal dialects among
the others, to the common Greek.” Thus on p. 47 [= Ravis 1656: 47]. Yet I would
prefer a mother–daughter relationship, on the one hand because we rightly claim
142      
Figure 10.3 Johann Heinrich Hottinger etched by Georg Meyer, 1664

the first place for Hebrew, just like a mother, and on the other also because some
deviate farther from Hebrew than that they can be called dialects. Among
dialects, there is equality of privilege. With the authority of the mother intact,
however, this cannot be given to the daughters. The dialects of the Greeks chiefly
consist in the change of letters, but in some aspects of (at least some) Oriental
languages there is such a great kinship with Hebrew as there is a greater
difference in other aspects. Arabic certainly differs from Hebrew in the number
and figure of letters, the number and form of endings, cases, the quantity of roots
(similarly Arabic is richer here than all others), the equipment of conjugations,
syntax, and formulas of speaking to such an extent that there is no dialect in
Greece that departs from the common Greek by so many steps as the number of
parasangs by which Arabic departs from Hebrew. It nevertheless does not cease
to be a daughter, as it very splendidly expresses in other respects the inborn face
of its mother, as becomes sufficiently evident from the lexicon itself.⁵
⁵ Hottinger (1661: a.3–4): ‘Hinc enim manifestissimè patet, . Linguam Sanctam reliquarum
omnium Orientalium esse matrem, cæteras filias, vel magis, vel minùs degeneres. Sunt, qui Dialecti
.    . . . 143
To summarize Hottinger’s somewhat complex account and its odd comparative

constructions: Aeolic, Doric, Attic, and Ionic could be seen as dialects of the Koine
Greek language, since they differed only by letter changes from it. In contrast, the
Oriental tongues were to be considered different daughter languages of one
common Hebrew mother, as they varied on many more levels than the letters
alone, including morphology, roots, syntax, and phraseology. Hottinger assumed
chronological primacy for Hebrew but not for the Greek Koine. In other words, he
justifiably separated the language/dialect distinction from the mother/daughter
contrast by comparing the Greek with the Oriental context. His subtle critique is
especially remarkable, since it stands in startling contrast with his blind adherence
to the age-old dogma of Hebrew primacy.⁶
Remarkable though Hottinger’s insights may be, they apparently did not exert
any influence of note on the theorizing about the conceptual pair by his contem-
poraries and immediate successors. The dominant trend was to define dialect as
opposed to language in synchronic terms, even though in actual practice many
scholars used these concepts in a very confused manner, conflating synchronic
meanings with the language-historical usage. Even if a few other early modern
criticasters could be cited, it is not until the modern era that one finds more
fundamental critiques of the distinction and the different interpretations attrib-
uted to it. That is, however, a matter for later chapters (see Part V).
10.3 A discursive strategy for historical classification
The Renaissance ushered in a new era in which the linguistic horizon of scholars
broadened remarkably (Chapter 4, Section 4.1). In order to cope with the seem-
ingly endless variation languages could exhibit, polyglot philologists started to
compare different languages and arrange them into groups, for which they
hic nomenclaturam amant. Christ[iano] Rav[io] Disput[atione] de Dudajim p. 29. Hebræa Chaldæa
Syra & Arabica Dialecti apparent uno numero lingua, quæ jam infaustè quorundam vitio discerptæ,
immane quantum dissidere creduntur, cum cohæreant, ut communi Græcæ Æolica, Dorica, Attica,
Jonica præcipuæ inter reliquas. Ita p. 47. sed malim matris ad filiam σχέσιν, tum quòd Hebrææ, ceu
matri, primas jure vendicemus, tum etiam, quòd longiùs ab Hebræa quædam abeant, quàm quæ
Dialecti vocari possint. Inter Dialectos ἰσοτιμία. At hæc, salva matris authoritate, filiabus dari non
potest. Græcorum dialecti in elementorum potissimum μεταβολῇ consistunt. At Orientalium
Linguarum (quarundam saltem) ita magna in quibusdam cum Hebræa est γειτνίασις, ut in reliquis
major sit differentia. Arabica certè ab Hebræa differt Elementorum numero & figura, terminationum
numero & forma, casibus, radicum quantitate (sic Arabica locupletior hic est, quàm reliquæ omnes)
conjugationum apparatu: constructione loquendi formulis, adeò ut nulla sit in Græcia Dialectus, quæ
tot passibus à communi recedat Græca, quot quidem parasangis ab Hebræa abit Arabica. Neque tamen
filia esse desinit, quòd in aliis nativum matris vultum, ceu manifestum satis ex ipso fit Lexico,
luculentissimè exprimat’.
⁶ On the ancient origin of the idea that Hebrew is the primeval language see e.g. Van Rooy (2013:
30–1); Denecker (2017: 63–4).
144      
designed various conceptual means and frameworks. This early modern com-
parative study of languages is today often termed, somewhat teleologically, pre-
comparative linguistics (Considine and Van Hal 2010: 63–5). One of the
conceptual means devised was the language/dialect distinction, which allowed
scholars to express various degrees of linguistic kinship and to classify language
forms into larger groups of different natures.
In the latter half of the sixteenth century, the language/dialect pair emerged as
one of the principal classificatory devices used by scholars trying to create order in
the vast linguistic diversity of the world, with a primary focus on European and so-
called Oriental tongues. An early landmark in this development was Conrad
Gessner’s Mithridates of 1555 (see Chapter 4). Joseph Justus Scaliger restyled the
distinction between language and dialect even more explicitly in historical terms,
as I have argued earlier in this chapter. Several other scholars discussed the
concept of dialect out of an interest in linguistic kinship and diversity, and even
more authors used language and dialect to express the affinity, or lack thereof,
between a wide range of languages. For instance, the Swedish philologist Georg
Stiernhielm (1671: f.1) applied the conceptual pair to tongues belonging to what
is now known as the Uralic family of languages. Sometimes a scholar felt obliged to
admit that there was not enough evidence to determine the precise status of a variety:
was it a language, a dialect, or even a subdialect? Comparing a number of Prussian
with Lithuanian words, the East-Prussian philologist Philipp Ruhig (1675–1749)
confessed in 1745 that the lexical sample available to him did not allow him to
determine the relationship between both tongues in more precise terms.⁷
The language/dialect pair proved to be a particularly welcome tool for early
modern scholars wanting to make sense of the poorly documented Germanic and
Celtic tongues anciently spoken in Germany and especially in Gaul. In this
context, they often made use of the phrase ‘to differ only in dialect’ (Chapter 5,
Section 5.5). For instance, the Reformed historian Caspar Peucer (1525–1602)
claimed that the late antique Goths, Franks, Burgundians, Longobards, Alans, and
others were all Germanic nations who spoke the same tongue, ‘as I do not
distinguish them in language, but in idiom and dialect only’.⁸ Scaliger’s colleague
Paulus Merula, in turn, discussed the linguistic situation of ancient Gaul as
follows:
On the language of the ancient Gauls, in as far as something certain can be stated,
it should primarily be noted what Caesar writes in the first book on the Gaulish
⁷ Ruhig (1745: 53): ‘Hierbey wird nicht verdrüßlich seyn, auch einige Preußische mit dem
Littauischen übereinkommende Wörter zu lesen, die Prætorius aus Grunovio anführet, welche aber
noch nicht ausmachen, ob diese beyde als dialecti, subdialecti, oder mit vernünftigen analogischen
Grammatischen Gründen von einander unterschiedene Sprachen seyn’.
⁸ Peucer (1562: 88–9): ‘Non enim lingua, sed idiomate tantum ac dialecto discerno Gottos,
Francos, Burgundos, Longobardos, Alanos, Sueuos, Anglos, sicut reipsa linguis non discreparunt’.
.    . . . 145
war, i.e. that the whole of Gaul is divided into three parts: one of these is
inhabited by the Aquitanians, another by those who are called Celts in their
own language but Gauls in Latin, and the third by the Belgians. All these differ
mutually in terms of language, habits, and laws. What that difference in language
was, whether it was in dialects only or in the entire language, so that the
Aquitanians, Celts, and Belgians each had a different language, has not yet
been definitively disclosed by anybody, at least as far as I know. After all, who
could state something sure and ascertained in a matter so obscure and remote
from our age?⁹
Although stressing the thorny nature of the issue, Merula (1605: 419–21) none-
theless did propose a conjecture himself. He opted for a difference on the level
of language. Yet all the Gauls spoke one language at first, Welsh, which the
Aquitanians later exchanged for Basque and the Belgians for a variety of
Germanic, with the Celts preserving their original Welsh language. Merula thus
presented a nuanced conjecture grounded in historical developments.
The manipulative usage of the conceptual pair and the phrase ‘to differ only in
dialect’ was not restricted to ancient Europe, since a similar strategy surfaced in
the work of scholars treating the linguistic diversity of the New World. It did so
most notably in a controversy between two Dutchmen, the geographer Johannes
De Laet (1581–1649) and the polymath Hugo Grotius (1583–1645). In this fierce
polemic, the issue of language occupied a major position (see e.g. Laes and Van
Houdt 2013). Grotius claimed that linguistic evidence proved a Scandinavian
descent of the American nations. De Laet, however, countered Grotius’s argument
by describing the vast linguistic diversity of New Spain as follows: their idioms ‘do
not only vary in dialects, but are entirely different languages’.¹⁰ One language
outshone all: Mexican, by which De Laet meant Nahuatl. He described its wide
spread and noticed that ‘it has begun to be common, just like the Latin language
once was in Europe and the Greek in Asia’.¹¹ De Laet was thus clearly putting the
conceptual pair to use so as to describe the linguistic situation in New Spain.
Earlier in his work, De Laet had expressed a similar view with regard to the
continent of America as a whole:
Another thing was the sheer infinite diversity not only of dialects, but even of
languages, too, and there was also such a large variation among neighbouring
⁹ Merula (1605: 419): ‘De Veterum Gallorum Lingua, quo certi quid statuatur, notandum in primis
quod Cæsar scribit li. . de Bello Gallico, totam Galliam in tris divisam partis; earumque Vnam incolere
Aquitanos; Aliam, qui ipsorum lingua Celtæ, Romana Galli appellantur; Tertiam Belgas: eos omnes Lingua,
Institutis, Legibus inter se differre. Quod in Lingua discrimen illud fuerit, an in Dialectis solum, an in tota
Lingua, ut ea alia fuerit Aquitanis, alia Celtis, alia Belgis, nemo adhuc, quod ego quidem sciam, diffinite
aperuit. Nam in re tam obscura tamque ab ævo nostro remota, quis certi quid & explorati statuerit?’
¹⁰ De Laet (1643: 27): ‘non tantum dialectis variant, sed plane diversæ sunt linguæ’.
¹¹ ‘communis esse cœpit, quemadmodum in Europa olim latina, & in Asia græca.’.
146      
peoples that they showed almost nothing in common among themselves and
differed from each other entirely in terms of customs and practices.¹²
In other words, the existence of many different dialects as well as languages in the
New World disproved, in De Laet’s eyes, a possible Scandinavian origin of the
American peoples.
10.4 Conclusion
During the late Renaissance, the conceptual pair language/dialect acquired a

language-historical dimension, apart from its various synchronic interpretations.
This new understanding had forerunners in sixteenth-century texts but gained
widespread acceptance only in the course of the seventeenth century, partly due to
the influence of Joseph Justus Scaliger and scholars in his network. The new
genealogical interpretation remained in vogue not only throughout the early
modern period, in spite of occasional criticism, but also in modern linguistics,
even up to this day (cf. Chapter 17).
Early modern scholars eagerly relied on the historical interpretation of the
language/dialect distinction and especially on variations on the newly coined
phrase ‘to differ only in dialect’ when trying to prove or refute the kinship
among two tongues and, by extension, among the peoples speaking them. It was
employed especially keenly in cases where there was room for speculation and
disagreement, as was the case with the poorly documented tongues of ancient
Europe such as Gaulish and with distant contemporary tongues such as the ones
in the New World. In other words, the less familiar an early modern scholar was
with a particular set of tongues, the more likely it was that he, wanting to give an
impression of the relationship among the tongues in question, resorted to the
conceptual pair as a welcome tool to describe or manipulate ethnic history in
agreement with his scholarly programme and underlying assumptions.
The diachronic interpretation must be framed within the historical-
comparative turn in early modern linguistic interests, which itself was, in part, a
side effect of the renewed sense of history current among humanist scholars. Or to
put it with the well-chosen words of Luce Giard, ‘the study of language partici-
pates in the general movement of historization impregnating Renaissance
thought’.¹³ Indeed, linguistic diversification was increasingly seen as a cyclical
historical process that took place gradually and into which the language/dialect
pair also fitted.
¹² De Laet (1643: 4): ‘Accedebat pene infinita non modo dialectorum, sed & linguarum ipsarum
diversitas, & inter vicinos quoque populos, tanta varietas, ut nihil fere inter se commune ostenderent;
moribusque & institutis toto cœlo inter se discreparent’.
¹³ Giard (1992: 221): ‘la linguistique participe du mouvement général d’historisation qui imprègne la
pensée de la Renaissance’.
11
Consolidation by elaboration
Drawing the balance
In the previous chapters, I have attempted to unravel how the language/dialect

pair became an established part of the metalinguistic apparatus of early modern
scholars. I have mainly done so by developing the idea that scholars elaborated
seven principal interpretations of dialect opposed to, and subsumed under, lan-
guage in the years 1550–1650. In this chapter, I want to draw the balance of how
scholars managed to consolidate the position of the conceptual pair in linguistic
thought by 1650.
11.1 The seven main interpretations: a synthesis
By way of summary I have listed in Table 11.1 the seven main interpretations of
the conceptual pair devised and established by scholars in the century between
1550 and 1650. They are summed up more or less in the order in which they can
be traced in the early modern source texts. It should be kept in mind that my
systematized analysis of the matter does not exactly mirror accounts of the time,
which tend to be not particularly organized and can be found in various contexts,
usually in grammatical, philological, historiographical, and ethno-geographical
works.
Not all interpretations were put into practice as diagnostic criteria to determine
whether a specific speech form was a language or a dialect. Whereas the
Aristotelian criterion and mutual intelligibility were often employed to this end,
the geographical and ethnic interpretations, rooted in traditional Greek defin-
itions of diálektos, were not. Instead, they served to simply explain what an author
took a dialect to be. Regularity, position in time, and status, although prominent in
definitions of the conceptual pair, were likewise only rarely put to explicit use as
criteria to set a dialect apart from a language. In fact, scholars often had precon-
ceived ideas about the status of a particular linguistic entity, which is why they
usually did not ponder at length on the language/dialect status of a speech form.
The seven interpretations could be variously combined, all the more since none of
them were mutually exclusive. Still, it is possible to group some of them together,
as I have done in the preceding chapters. Not unsurprisingly, geographical and
ethnic coverage often co-occurred, as they were both inspired by Greek definitions
DOI: 10.1093/oso/9780198845713.001.0001
148   
Table 11.1 Summary of the main early modern interpretations of the conceptual pair
Interpretation Language Dialect
Nature of linguistic Substantial Accidental

differences (Aristotelian
criterion)
Geographical coverage Extended Restricted
Mutual intelligibility (Generally) absent among (Generally) present among
languages the dialects of a language
Regularity Regulated, analogical norm Anomalous deviation
Ethnic(-political) Group of nations One nation
coverage Also: state Also: part of the state
Position in time Primary Secondary
(language-historical
conception)
Status Superior, pure, literary Inferior, corrupted, unliterary
~ higher social classes ~ lower social classes
and can be regarded as two sides of the same coin: a dialect was spoken by a
restricted group of people living in a narrowly delimited area. Further, mutual
intelligibility can be regarded as a consequence of the Aristotelian criterion. These
interpretations were indeed mentioned together by several scholars. What is more,
numerous authors were aware that both features are gradual rather than strictly
binary. Regularity and status, too, were closely linked and often even merged in
the seventeenth and eighteenth centuries, even though not all scholars bracketed
them together. This association is especially rare in cases where the literary Greek
dialects occupied a prominent place. They were usually perceived as deviating
from the analogical Koine, without, however, being stigmatized as inferior lin-
guistic entities, as many vernacular dialects were.
Somewhat peculiar was the language-historical interpretation of the conceptual
pair, especially since this conception was the only one approaching the relation of
dialect to language from a diachronic rather than a synchronic perspective. It was
nevertheless not uncommon for later scholars to intuitively link it to the inter-
pretations in terms of regularity and status, assuming that the analogical and
superior norm, the language, was inherently prior to anomalous and inferior
deviations from it, its dialects. The Dutch language scholar Lambert ten Kate
(1674–1731), for instance, juxtaposed language to dialect synchronically in
normative terms as well as diachronically in a genealogical sense but did not
problematize this double usage in his otherwise highly detailed Introduction to the
knowledge of the sublime part of the Dutch language. Ten Kate (1723: .57–8) spoke
at one instance of the different Dutch dialects, which he distinguished from the
common written language in his book. Elsewhere, however, he discussed
.     . . . 149
pronunciation rules across different dialects, in which case he meant the historical
dialects of the Germanic branch of tongues (ten Kate 1723: .165). Such diverging
usage is further evidence for my hypothesis that early modern authors perceived the
conceptual pair mainly as a flexible metalinguistic tool that need not be questioned.
The interpretations of the conceptual pair emerging in the early modern period
can be variously related to the Greek tradition, contemporary scholarly interests,
and the sociolinguistic realities of Western Europe, a dimensional triangle visual-
ized in Figure 11.1. Some understandings were inspired by traditional Greek
definitions of diálektos, such as the interpretations in terms of geographic and
ethnic coverage, or indirectly motivated by Renaissance grammars of Greek or
Greek philosophical categories, as in the case of the Aristotelian criterion. Most
interpretations were, however, primarily the product of early modern scholarly
interests, even though they were often presented as ahistorical givens rather than
as innovations, much like the conceptual pair itself. The language-historical
interpretation in particular should be situated close to this pole, since its emer-
gence was tightly connected to the emerging early modern interest in language
history and diversity. Certain meanings of the language/dialect distinction, espe-
cially that of status, reflected first and foremost evolutions in the sociolinguistic
realities of early modern society and were for a large part shaped by them. These
circumstances included most notably standardization processes, the fact that the
standard language spread very unevenly across different social classes, and the
increasing importance of linguistic borders in the constitution of political states.
Put differently, interpretations of the conceptual pair displayed the typical early
modern blend of old and new, albeit in this case with a preponderance of new
elements, even though scholars usually did not present them as such.
Greek tradition
Figure 11.1 The dimensional triangle

shaping the early modern conceptual Sociolinguistic Scholarly
realities interests
pair
One could say that the appropriation of the Greek term diálektos and its
interpretations was more or less completed in the course of the seventeenth
century, a process not merely consisting in passive reception but displaying a
major creative dimension. The word was, on the one hand, borrowed from Greek
into Neo-Latin and most Western European standard languages. On the other
hand and more importantly, the appropriation consisted in adding new meanings
150   
to it. This semantic revolution primarily happened by the opposition of the term
to words meaning ‘language’ and the resultant various interpretations of this
metalinguistic distinction. One should, however, differentiate between present-
day perspectives and those of the early modern appropriators of the concept. My
analysis suggests that early modern scholars made considerable changes to the
original meanings of diálektos, gradually disassociating the word from the Greek
heritage. Even so, most of them seem to have assumed that they were simply
adopting a Greek term without changing its semantics too much, if at all
(Chapter 6, Section 6.5).
Early sixteenth-century scholars had left implicit how they conceptualized a
language as opposed to a dialect, even though it can be inferred from their work
that they often equated language with the standard variety being designed. In the
late sixteenth and seventeenth centuries, scholars started to be more explicit about
the ways in which they understood the language pole. It may suffice to quote what
the famed grammarian of German, Justus Georg Schottel, had to say on the
language variety he was describing:
The High German language, however, which we discuss and at which this book
aims, is not a dialect properly speaking, but the German language in its essence,
as the learned, wise, and expert men have finally regained it and use it.¹
This conception was projected back onto Greek as well. Tellingly, the Dutch
Calvinist theologian and orientalist Johannes Leusden (1624–99) preferred to
call the Greek Koine ‘the language in its very essence’ (lingua ipsissima) rather
than ‘common dialect’ (dialectus communis; 1670: 86). The idea of a common
language did not, however, become the only interpretation of the language
concept. On the contrary, it seems wiser to regard the early modern language/
dialect distinction as comprising several conceptual oppositions, most import-
antly common (standard) language vs. dialect and language mother vs. derived
dialect, but more marginally also language as a linguistic conglomerate consisting
of several related dialects.
It would therefore be misleading to speak of the early modern language/dialect
pair, as if it were a fixed and universally accepted opposition, especially since
several interpretations could be applied to it. It is for this reason perhaps better to
think of the conceptual distinction as a kind of flexible matrix into which
diverging and context-dependent meanings could be fitted. Its interpretations
depended on various circumstances, such as, for instance, the linguistic context
on which a scholar was commenting, his discursive aims, and his intellectual
¹ Schottel (1663: 174): ‘Die Hochteutsche Sprache aber/ davon wir handelen und worauff dieses
Buch zielet/ ist nicht ein Dialectus eigentlich/ sondern Lingua ipsa Germanica, sicut viri docti, sapientes
& periti eam tandem receperunt & usurpant’.
.     . . . 151
outlook and underlying assumptions. What is more, one single scholar could
make varying use of different interpretations of the conceptual pair, as the case of
Lambert ten Kate cited above served to illustrate.
I must note here that the interpretations I have distinguished for the early
modern language/dialect pair have been distilled from various works of dif-
ferent genres, written by a wide range of scholars with diverging backgrounds
and interests. These texts can nevertheless be categorized into two main
groups. Firstly, there were a restricted number of scholars such as Georg
Stiernhielm, who had an obvious and engaged interest in the distinction,
although always subordinate to a higher, usually grammatical, philological,
or historiographic goal. The case of Stiernhielm for this reason deserves a
more thorough discussion, which it will be accorded in the next and final
chapter of the third part of this book. Secondly, there was a large group of
authors who were not actively involved with the interpretation of the concep-
tual pair, but whose works do contain revealing testimonies offering traces,
however faint, of early modern ideas about the language/dialect distinction.
Antoine Furetière’s dictionary definition of dialecte may count as an example
of the latter group. Furetière did not offer an interpretation of the word
because he had any special interest in it, but because he was trying to compile a
dictionary of French words, and as it happened, dialecte had become a naturalized
French word.
In the years 1550–1650 the language/dialect distinction was fleshed out in a
more straightforward and explicit manner than in the first half of the sixteenth
century. This development continued during the last part of the early modern
period, even though no totally new interpretations of the conceptual pair were
proposed. As I argue in Part IV, the contribution of the period 1650–1800 lies
elsewhere.
11.2 Emancipating the conceptual pair

from the Greek heritage
I have mentioned that only two interpretations of the conceptual pair can be
directly traced back to Greek scholarship on the dialects, those emphasizing the
limited ethnic and geographic coverage of a dialect vis-à-vis the language on
which it was claimed to depend. This low level of indebtedness to Greek sources
might suggest that in the consolidation period the concept of dialect was gradually
emancipated from the Greek heritage and the study thereof. Did this emancipation
indeed take place? And if so, what other indications of this tendency can be
found? The most compelling evidence in favour of this hypothesis seems to be the
introduction of two major interpretations of the language/dialect pair: the analysis
in terms of status and the language-historical sense.
152   
The strongest indication of the emancipation from the Greek heritage is the
seventeenth-century view that a dialect was a variety of speech corrupted from a
language (cf. Burke 2013: 29–30). I have mentioned that the renowned German
language scholar Justus Georg Schottel claimed that ‘in all dialects there is
something faulty, which cannot occupy the position of rule in the language itself ’
and how Antoine Furetière defined dialecte as ‘particular tongue of a province,
corrupted from the general or principal language of the kingdom’ (Chapter 9,
Sections 9.1 and 9.3). Even though Furetière also referred to the Greek dialects, his
interpretation of the term was clearly grounded in his native French context. This
French anchoring emerges not only from the fact that he considered a dialect to be
a corrupted form of speech but also from his mentioning a kingdom, which
cannot be anything other than a reference to the form of government of France
in his age. The idea that a dialect was a corrupted form of a language was almost
entirely absent from earlier thought, no doubt because the traditional Greek
dialects were not conceived as despicable speech forms. On the contrary, they
were the linguistic media of a revered literary tradition.
The emancipatory tendency emerges also from the work of scholars operating
with the language-historical sense. A key figure here was, as I have established, the
philologist Joseph Justus Scaliger. For his interpretation of dialect as an offshoot of
a mother language, Scaliger was inspired neither by Ancient Greek dialectal
diversity nor by Greek theorizing. What is more, Scaliger’s one-sentence discus-
sion of the Greek mother language left the reader in doubt as to which episode in
the history of the Greek language he was referring to:
There are several idioms of the ΘΕΟΣ mother [sc. the Greek language; theós
(θεός) means ‘god’ in Greek], and this is not astonishing with so many distant
islands, which differ very much in terms of location as well as the interaction of
language.²
Was Scaliger alluding here to diversity in the ancient or the contemporary

vernacular language? The fact that he did not mention the canonical Ancient
Greek dialects indicates that the last option is highly likely. This conjecture is
further supported by his statement that he was taking into account eleven
‘language mothers’, since ‘today there are not more of them extant in the whole
of Europe’.³ In other words, Scaliger operated with the distinction between a
language mother and the dialects derived from it without mentioning the proto-
typical Ancient Greek context or relying on Greek scholarship. His silence most
² Scaliger (1610a: 121): ‘Matricis ΘΕΟΣ, pluria sunt Idiomata, quod non mirum in tot Insularum
interuallis, quæ vt loco, ita linguæ commercio valde dissident’.
³ Scaliger (1610a: 119): ‘Sunto igitur nobis Matrices eæ, quæ per omnia inter se discrepant,
cuiusmodi , non amplius hodie supersunt in vniuersa Europa’ [emphasis mine].
.     . . . 153
certainly cannot be explained by a superficial knowledge of Ancient Greek on

Scaliger’s part. In fact, he was an excellent Hellenist—he claims to have learned the
language in only twenty-one days by reading Homer (Scaliger 1594: 56)—and
evinced interest in the Greek dialects in other parts of his oeuvre (Grafton 1983:
102–4, 155). He was even creative with the language, coining a new Greek term to
denote a variety subsumed under a dialectus, which we might call a subdialect:
paradiastrophē ́ (παραδιαστροϕή), a word that does not occur in extant Ancient
Greek and Byzantine texts (Scaliger 1610b: 123). Scaliger’s case suggests that
trained Hellenists, too, started to omit references to the Ancient Greek dialects
and Greek scholarship where one would expect them to be present.
The emancipation also surfaces from the tendency to provide idiosyncratic
definitions of the term dialect, even in handbooks and dissertations on the Greek
language and its dialects, which one might suspect to be rather conservative in this
regard and which depended on Greek scholarship in many other respects.
A revealing case in point is Erasmus Schmidt’s analysis of the word dialectus,
part of which I have already discussed in Chapter 7. Schmidt enumerated four
different meanings. For the first three, he referred to Greek texts. He mentioned
Aristotle’s peculiar definition ‘articulation of the voice by means of the tongue’,
the sense of ‘conversation’ present in Plato’s dialogues, and the meaning of
‘language entirely different from other languages’, for which he referred to the
New Testament, Acts 2.6 and 2.8 in particular. Schmidt did not, however, use the
word in these senses, as he himself indicated. He offered a fourth interpretation
instead, for which he did not rely on Greek scholarship. His own definition
unmistakably made a distinction between a dialect and a language, which was
motivated by his assumptions about vernacular German and which he projected
back onto the Ancient Greek language:
Lastly, it sometimes denotes, like in the present treatise, a speech particularity of

some one and the same language, which is different in the different provinces and
cities using this language. For instance, the Saxons, Misnians, Thuringians,
Silesians, Franconians, Swabians, etc. all use the German language; and still
each of them has their own dialect, i.e. a peculiar way of writing, pronouncing,
forming, constructing, or even naming as a whole.⁴
In contrast to Schmidt, the French Protestant scholar Matthieu Cottière

(1581–1656) claimed that the original polysemy of the Greek word diálektos had
⁴ Schmidt (1604: 3–4): ‘Aliquando denique, ut in præsenti tractatu, notat idioma unius & eiusdem
alicuius linguæ quod diversis Provinciis vel Civitatibus eâ utentibus, est diversum. Exempli gratia:
Germanica lingua omnes utuntur, Saxones, Misnici, Thuringi, Silesii, Franci, Sueci, &c: & tame[n]
singulis sua est Διάλεκτος, hoc est peculiaris, vel scribendi, vel pronunciandi, vel formandi, vel
construendi, vel planè etiam nominandi ratio’. In the early modern period, Sueci usually denoted
‘Swedes’, but Schmidt is more likely to have meant the Swabians here (Suevi/Suebi).
154   
disappeared in the early modern period. Whereas the word designated both
‘language’ and ‘a certain affection of a language’ in Greek texts, it currently only
conveyed the latter meaning, Cottière (1646: 208–15) contended. He thus undeni-
ably displayed an awareness that there was a disparity between Greek and early
modern usage, which makes him rather exceptional, since most scholars silently
assumed conceptual continuity. However, like many of his contemporaries, he was
reading the subsumption of dialect under language into the Greek sources on
which he was relying.
Some scholars clung to the Greek tradition and the Ancient Greek linguistic
context. For instance, according to certain authors, the term dialect could exclu-
sively refer to a variety of the Greek language. As a matter of fact, the Ancient
Greek context remained the prototypical point of reference in discussions of the
concept of dialect, even though it was, in its subsumption under language, an early
modern innovation. What is more, the tendency to employ Greek as a reference
context also holds true for more obviously new concepts such as standard
language, which was occasionally projected back, anachronistically, onto the
Ancient Greek Koine.
A great friction between the Greek heritage and early modern conceptualiza-
tions is observable in several cases. A number of scholars even formulated
idiosyncratic definitions of diálektos in Ancient Greek. By proceeding thus, they
wanted to convey the impression that they were offering a received interpretation
taken from Greek sources, even though they were in fact offering definitions of
their own, which they had adopted from the Greek tradition and attuned to their
own assumptions. A remarkable case in point is the Bruges-born scholar Andreas
Hoius, professor of Greek at Douai university. Hoius provided his readers with the
following definition, expressed partly in Greek:
we call the character of the language with its own form of speaking dialect . . . Or
to explain it in rather rough terms, a certain property of speech, by which a
variety of speaking is distinguished, which usually and commonly exists between
different regions of the same nation.⁵
As I have already argued (Chapter 7, Section 7.1), the second part of Hoius’s
definition was, with slight adaptations, silently taken from an earlier Renaissance
grammarian, the French Hellenist Petrus Antesignanus. The first part, however,
appears to have been a product of Hoius’s own mind. Inspired by traditional
Greek views, he formulated a definition presupposing a subsumption of dialect
⁵ Hoius (1620: 95): ‘Διάλεκτον dicimus τὸν τῆς γλώττης χαρακτῆρα, ἴδιον ἔχοντα τοῦ λέγειν σχῆμα,
linguæ characterem, qui propriam habeat loquendi formulam. Sive, ut παχυμερέστερον explanemus,
Sermonis quandam proprietatem, qua loquendi varietas distinguitur, quę inter diversos eiusdem
nationis tractus plerumque solet existere’.
.     . . . 155
under language—especially when read in connection with the subsequent quote

from Antesignanus’s work. In sum, Hoius offered an idiosyncratic definition of
the term diálektos, but he did so in Greek so as to keep up the appearance of
traditionalism.
I have already established how Antoine Furetière’s conceptualization of dialect
as a speech form corrupted from a language was influenced by his native French
context. It was unproblematic for Furetière to extrapolate this idea to the Ancient
Greek context, even though the traditional Greek dialects were normally not seen
as corruptions of the Greek language in the early modern era. This uncommon
view might be related to a lack of acquaintance with the Greek context on
Furetière’s part and, perhaps, to a scepticism about Greek culture in general.
Indeed, the French lexicographer elsewhere stressed the superiority of contem-
porary over Greek philosophy (Stanton and Wilkin 2010: 30). However, even
professional Hellenists struggled with the tension between the Greek tradition and
their own conceptualizations and assumptions. For example, in a 1684 disserta-
tion presented in Wittenberg, although stressing the similarities between Ancient
Greek and German dialects, two German Hellenists associated the dialects of their
native tongue with lower social classes (Kirchmaier and Crusius 1684: .2v–3).
This connection contrasted with their assumption that the Greek dialects were
esteemed literary varieties.
In summary, there was a general tendency towards emancipating the dialect
concept from the Greek heritage. Nevertheless, seventeenth-century authors
often still associated dialectal variation closely with the Ancient Greek lan-
guage. Furthermore, scholars frequently interpreted the Greek context by
means of the language/dialect distinction. This analysis was, however, a pro-
jection of their own ideas onto a context distant in time and space from their
own, betraying the assumption that their native vernacular dialectal diversity
was comparable to Ancient Greek variation (cf. Van Rooy 2020a: Chapter 8).
Additionally, early modern scholars elaborated on definitions from the Greek
tradition, into which they silently introduced a distinction between language
and dialect themselves. Put differently, even though theorizing on the concep-
tual pair increasingly stepped out of the shadow of the Greek heritage in the
period 1550–1650, this emancipation often occurred in a highly ambiguous
manner. Also, the Ancient Greek dialects continued to attract most scholarly
attention throughout the seventeenth and eighteenth centuries for philological
reasons.
To gain better insight into the emancipation from the Greek heritage and the
ambiguity of this process, it might be fruitful to resort here again to a process
I have introduced in Chapter 8, Robert K. Merton’s (1968: 27–8, 35) obliteration
by incorporation. Indeed, the de-Hellenization seems to display two main
properties of this process. Firstly, scholars increasingly omitted references to
the Greek sources in their definitions of dialect, even though these often
156   
constituted the bases for their own interpretations. This independence contrasts
with sixteenth-century practice, when scholars still either cited a specific Greek
author or generically referred to ‘the Greeks’ in their definitions. Secondly,
obliteration by incorporation is most likely to take place when the original
contributors are renowned, which may also be said to hold true for the
Ancient Greeks as a people, held in high esteem by Renaissance humanists.
Applying Merton’s concept here has disadvantages, too. Most importantly, it
presupposes modern methodological practice and therefore does injustice to
certain aspects of early modern scholarly practice, in which matters such as
originality and transparency in methodological and conceptual terms were not
always prime concerns. Also, unlike the sciences with which Merton was
concerned, the process of emancipation I have been describing here did not
involve an autonomous field of study. Still, it seems useful to tie the de-
Hellenization of the dialect concept to broader tendencies in the history of
ideas, so as to arrive at a better understanding of this conspicuous transform-
ation of the language/dialect distinction.
11.3 Conclusion
The conceptual pair underwent two main transformations in the hundred years
between about 1550 and 1650: a consolidation by elaboration of its interpretations
and an emancipation from the Greek heritage. Even though I have opted to
investigate these two evolutions somewhat distinctly with a focus on the first
development, I should point out here that they often went hand in hand. Their
compatibility can be neatly exemplified by repeating here the dialect lemma in
Thomas Blount’s Glossographia, an influential English dictionary from 1656:
Dialect (dialectus) is a manner of speech peculiar to some part of a Country or

people, and differing from the manner used by other parts or people, yet all using
the same Radical Language, for the main or substance of it.
(Blount 1656: s.v. dialect; cf. already Chapter 8, Section 8.1)
Blount, on the one hand, provided a definition of his own, which combined several
interpretations: the Aristotelian criterion as well as geographical and ethnic
coverage. The Greek background was, on the other hand, not prominently
present. Blount referred to the Ancient Greek dialects only after mentioning
variation in English. What is more, he was entirely dependent on the geographer
Peter Heylyn (1599–1662) for information on the Greek dialects. Blount’s definition
suggests that scholars of Greek language and literature had lost their initial mon-
opoly over the conceptual pair, with non-Hellenists increasingly contributing to its
interpretations.
.  157
A major consequence of the de-Hellenization was that the seventeenth century

brought an ever-increasing emphasis on the secondary nature of a dialect vis-à-vis
a language in functional, grammatical, chronological, and sociocultural terms
even if the etymological value of dialects was often stressed. In other words, a
dialect was more and more regarded as deviating from, as well as inferior and
posterior to, a language. Consider, for instance, the view, expressed in an early
seventeenth-century dissertation On languages presented in Leipzig, that the
Germanic language was ‘a dialect of none’ (‘nullius dialectus’), an assertion
taken to imply correctness, purity, and antiquity (Bachmann and Ludovici
1625: .1). A negative interpretation of the dialect concept had initially been
blocked by its close association with the admired literary Greek dialects, which
is, however, not the entire explanation. The shift was also connected to the rise
in prestige of the vernacular languages as a result of intensifying standardization
processes. Normative varieties received an advanced grammatical codification
and were ever more widely accepted and used, especially in France, where this
evolution took place at the quickest pace. Linguistic standardization was encour-
aged by political centralization in unified states such as England and France in
particular, as I have put forward in Chapter 7, Section 7.5. In the process, the
conceptual pair sometimes received a political colour in these areas already
during the seventeenth century.
In more fragmented regions, too, scholars sometimes assigned a political
meaning to the language/dialect distinction, albeit less prominently so. The
German orientalist Hiob Ludolf (1624–1704), notably, criticized his former friend
Andreas Müller (1630–94) for having included specimens not only of languages
but also of dialects in his Lord’s Prayer collection of 1680. Interestingly, Ludolf
alluded to the way in which one ought to distinguish between the two concepts in
his manuscript notes, written down in his interleaved copy of Müller’s book:
Whatever tongues only differ as to provincial pronunciation, are not found in

any printed book whatsoever, and do not have their own king or lord, these do
not deserve to be listed, as, for instance, Guelderish, which does not differ from
Dutch, unless in terms of vulgar pronunciation.⁶
Ludolf put his ideas into practice and struck out certain specimens, for example of
Swiss German, which he saw not as a language but as a dialect. The German
scholar seems to have taken the conceptual pair principally in two senses here: the
Aristotelian interpretation and the language-external criterion of cultural and
political status. As opposed to languages, dialects differed from each other only
⁶ ‘Quaecumque linguae tantum prouinciali pronuntiatione differunt, nec in ullo aliquo libro
impresso reperiuntur, nec suum proprium regem aut herum habent, eae non merentur referri ex. gr.
Geldrica, quae non differt a Belgica nisi plebis pronunciatione’ (quoted from Van Hal forthcoming).
158   
superficially, were not used in writing, and could not invoke the authority of a
political entity. Ludolf ’s suggestion that dialects ‘only differ as to provincial
pronunciation’ may moreover imply that he presupposed mutual intelligibility
among related dialects but not among languages.
In sum, some seventeenth-century scholars politicized the conceptual pair to a
certain extent. In doing so, they paved the way for the modern quip, ‘a language is
a dialect with an army and navy’. However, I must agree with Peter Burke (2004: 7)
that the politicization of both concepts remained a marginal phenomenon in the
early modern period. Indeed, as I suggest in Part V (especially Chapter 23), this
interpretation rose to prominence only during the nineteenth century. Before
treating this major evolution, however, I have to turn my attention to the fate of
the conceptual pair in the last 150 years of the early modern period and, to begin
with, to the transitional figure of Georg Stiernhielm.
12
The conceptual pair in transition
The case of Georg Stiernhielm
The previous chapters have offered a synthetic account of the consolidation of

the conceptual pair in the years 1550–1650. My choice of presentation has to a
considerable extent been determined by the available evidence, which itself is
rather fragmentary, since before 1650 few scholars reflected extensively on the
distinction between language and dialect. After around 1650, certain authors
started to provide more systematic treatments of the issue against the backdrop
of philology and historiography. An early witness to this change is the Swedish
philologist Georg Stiernhielm (Figure 12.1), whom I have briefly mentioned
already several times (Chapter 8; Chapter 10, Section 10.3). His views are, how-
ever, remarkable enough to deserve a separate discussion in a historical overview
of the conceptual pair, all the more since he marks the transition of the consoli-
dation period (Part III) to the systematization period (Part IV).
12.1 Georg Stiernhielm, a Swedish all-round scholar
Born as Göran (Georg) Olofsson in a small Swedish village in the Dalarna region
on 7 August 1598, Georg Stiernhielm carved out a brilliant career in different
branches.¹ After studies in Uppsala, Greifswald, and elsewhere, and after travels
across Europe, Stiernhielm put his expertise and knowledge into practice by
serving the Swedish state in many different capacities, mainly as a jurist but
also as a soldier and archivist. A child of middle-class parents, he was able to
acquire the status of nobility in 1631, taking on his new name Stiernhielm,
meaning ‘star helmet’, to match this step forward. Although a man of many
talents and despite his new status, he faced financial problems for almost his
entire life, up until his appointment as director of the College of Antiquity
(Antikvitetskollegium) in 1667, five years before his death on 22 April 1672. The
saying on his tombstone ‘Vixit, dum vixit, laetus’—‘As long as he lived, he lived
happily’—is therefore somewhat surprising. His poverty might not be unrelated
¹ On Stiernhielm’s life almost all writings available are in Swedish. See e.g. Swartling (1909), who
discusses Stiernhielm’s scholarly activities at length.
DOI: 10.1093/oso/9780198845713.001.0001
160     
Figure 12.1 Georg Stiernhielm

to his supreme arrogance and fickle character. He was, for instance, known to
have incurred a nasty injury to his right arm after a bar fight in Dorpat, present-
day Tartu, Estonia, in 1641. The damage he suffered was lasting and he was
forced to write with his left hand for the rest of his life. Moreover, when in 1649
Stiernhielm returned to Stockholm after a military debacle and became part of
the intellectual circle around Queen Christina of Sweden (1626–89), he fell out
of grace because he insulted some of the queen’s confidants.
Stiernhielm’s volatile character has, however, not influenced the reception of
his work as a writer, especially his poetical oeuvre, for which he is mainly known
today. What is more, Stiernhielm is even hailed as the father of Swedish poetry for
his Hercules, a national epic composed in classical hexameters adapted to the
stress accent of his native tongue. Stiernhielm, however, made a name for himself
as a scholar, too, in the field of, among other things, mathematics and language
studies. The wide range of his interests meant, however, that he did not achieve any
fundamental contributions in most branches, and for this reason a nineteenth-
century compatriot of his, Erik Gustaf Geijer (1783–1847) even labelled him ‘a great
.  ,  -  161
beginner in everything’.² This judgement might be too harsh, since his contribution
to philology and language studies seems to go beyond the level of a beginner,
even a great one. His originality can, for instance, be easily recognized, when
one reads two lengthy passages from his pen, which shed light on his concep-
tualization of the language/dialect pair:
Here, we must first know how a language differs from a dialect. Languages
mutually differ in substance, in their foundation as it were, but dialects differ
in accident. Languages are distinguished by matter, dialects by form. And, to put
it more plainly, every language is defined by its own words and roots of words,
which are both unknown and foreign to another language. A dialect is a deflec-
tion of one language into a separate idiom of a certain people, which does not
differ by the identity of roots and words, but through the formation of accidents,
cases, endings, and the addition, detraction, transposition, and change of a letter
or syllable in some words. To these features are added the often different
composition and usage of the same words, the variation of accent and spiritus,
and above all a pronunciation departing from the common one.³
I call dialect a notable difference of an idiom departing from a first language into
different directions. And this difference certainly consists in the occasionally
diverging usage of substantial and radical words and in their corruption to
another meaning, in their composition, phrasing, declension, conjugation, pro-
nunciation, and accent. If these elements, I say, are so varied that they constitute
a remarkable difference but are even so understood, albeit in some cases with
difficulty, they constitute a new dialect, even though it was not mixed up with any
other foreign language. And for this reason, all languages have their peculiar
dialects. Italian has Tuscan, Lombard, Venetian, and Neapolitan; French has
Parisian, Toulousian, Picard, Norman, and Provençal; do German, Swedish, and
so on not have so many dialects as they have provinces? Every language has its
varieties. Yet I regard as different those languages that mutually differ by such a
distance that the substantial words themselves and the accidental forms seem
entirely different, to such a degree that, in general, those mutually conversing
cannot understand each other in any way, unless through an interpreter. Dialects
become like this over a long span of time, while mutually departing from each
² Geijer (1850: 160): ‘i allt en stor börjare’.

³ Stiernhielm (1671: c.4): ‘Hic, Lingua, quo differat à Dialecto, prænoscendum est. Linguæ inter se
substantia, ceu subjecto; Dialecti vero Accidenti differunt. Linguæ Materia, Dialecti Forma distingvun-
tur. Et ut explicatius dicam: Lingua quæq[ue] vocabulis & ipsis vocabulorum radicibus sibi propriis, &
alienæ linguæ incognitis & peregrinis, definitur. Dialectus est unius linguæ deflexus, in singulare
Nationis alicuius idioma, radicum & vocabulorum identitate non differens, sed formatione acciden-
tium, casuum, terminationum, litteræ aut syllabæ in quibusdam additione, exemptione, trajectione,
mutatione, quibus accedit compositio, & usus earundem vocum sæpe diversus: Accentus ac Spiritus
variatio, & super omnia, à communi recedens pronuntiatio’.
162     
other further and further, a process in which, as with all things natural, the
character of natural progression necessarily appears. As a matter of fact, two
entities that were one, when deviating from their unity, the further they evolve
and the more borders they cross, the greater the interval is by which they differ
both from the unity, their origin, and mutually from each other.⁴
Stiernhielm introduced both original definitions of the conceptual pair in the

preface to his edition of bishop Ulfilas’s fourth-century Gothic New Testament
translation, published in 1671, one year before his death. He had, however, already
drafted large parts of this text around the years 1651–3, at which time he
conceived it as part of a planned work entitled Runa Suethica, an outline of
which appeared in print in 1669.⁵ Stiernhielm’s preface treated the origin of
languages and provided an outline of his language-historical programme, in
which the conceptual pair played a key auxiliary role. It served as a handy tool
for determining the historical position of different language varieties and their
degree of kinship. More specifically, Stiernhielm used it to relegate the sacred
language Hebrew to the status of dialect within the family of so-called Oriental
tongues. As mentioned in Chapter 10 (Section 10.2), Hebrew was widely assumed
to be Man’s original tongue before the modern period, which was why Stiernhielm
(1671: c.3–d.4) felt compelled to defend his view at great length. Another main
thesis of Stiernhielm was that his native Swedish was ‘the least altered represen-
tative of the ancient Scythian protolanguage’ (Considine 2008a: 308), revealing a
clear patriotic bias on the part of Sweden’s first national poet. Stiernhielm thus
showed himself to be a follower of the so-called Scythian hypothesis by suggesting
that Swedish was a dialect of Scythian. This theory had received different inter-
pretations in the early modern period, but the core assumption was more or less
the same in all of its varieties: Scythian was the now lost protolanguage of most
⁴ Stiernhielm (1671: d.4–e.1): ‘Dialectum voco, idiomatis à prima lingua in diversa abeuntis,
insignem differentiam. Quæ quidem differentia consistit in vocum substantialium seu radicalium
diverso interdùm usu, & ad aliam significationem depravatione, compositione, phrasi, declinatione,
conjugatione, pronunciatione & Accentu. Hæc, inquam, ita variata, ut insignem constituant differ-
entiam, nec tamen secus intelligantur, licet in quibusdam ægrè, novam constituunt dialectum, etiamsi
nulli aliæ linguæ exoticæ commixta fuerit. Atq́[ue] ea ratione, linguæ omnes suas habent peculiares
dialectos. Italica Thuscanam, Lombardicam, Venetam, Neapolitanam; Gallica Parisiensem,
Tholosanam, Picardicam, Nortmannicam, Provincialem; Germanica, Svedica, & quæ non tot habent
Dialectos, quod [sic pro quot] habet Provincias? singulæ suas habent varietates. Linguas verò pro
diversis habeo eas, quę eo inter se intervallo distant, ut ipsæ voces substantiales, & forma Accidentalis
prorsus aliena videantur, adeò ut vulgò inter se colloquentes, nullo modo, nisi per interpretem, mutuo
se intelligere queant. Tales evadunt dialecti longâ temporum morâ, longius longiusq́ue à se invicem
recedentes, in quo, ut naturalia omnia, necessum est, ut progressionis naturalis indoles elucescat. Duo
enim quæ unum fuere, ex unitate sua digredientia, quò longius & pluribus terminis progrediuntur, eò
majori intercapedine, & ab unitate, primordio suo, & à se invicem, distant’. Eskhult’s (forthcoming b)
English translation loosely inspired my renderings.
⁵ See Eskhult (forthcoming b) for an edition and English translation of the preface, taking into
account manuscript testimonies. The Runa Suethica outline is Stiernhielm (1669).
.       163
tongues of Europe and some Asian languages. The hypothesis thus anticipated the
nineteenth-century concept of Proto-Indo-European in a number of respects.⁶
12.2 The conceptual pair according to Stiernhielm
With the position of the language/dialect distinction in Stiernhielm’s language-

historical investigations in mind, I can now look in more detail at the ways in
which he interpreted it.⁷ Reading his two definitions and other parts of his
writings, the reader discovers one essential difference from his predecessors and
contemporaries that stands out immediately: the length and detail of Stiernhielm’s
treatment of the conceptual pair. Furthermore, his main criterion for distinguish-
ing between language and dialect was absent from sixteenth-century definitions.
Languages, he argued, varied in substance, on the level of the lexicon and roots,
whereas dialects differed accidentally, on the level of accent, pronunciation, letter
mutations, morphology, the meaning of words, and phraseology. The latter kind
of variation Stiernhielm (1671: e.2) compared to the different looks individual
persons can have. I have identified this criterion as the Aristotelian interpretation
in Chapter 8, for which Stiernhielm served as the introductory example. It is not
this common understanding of the conceptual pair that deserves attention here,
but the fact that he explicitly linked both poles to different linguistic features.
Most of the features mentioned are probably well-known to modern readers,
but one might call for further explanation. In the first passage Stiernhielm argued
that related dialects differed by ‘the addition, detraction, transposition, and
change of a letter or syllable in some words’. A long history lies behind this
remark. Since Roman antiquity language scholars had identified the four main
letter-change processes mentioned by Stiernhielm, which were usually known
under the Latin phrase permutatio litterarum, ‘permutation of letters’ (Chapter 4,
Section 4.4). Letter-change processes were used very flexibly to various ends, usually
to formulate speculative etymologies of words or in the early modern period
often also to prove the alleged kinship between two tongues. Stiernhielm had,
however, a particular use for these letter changes, since he associated them with
differences on the level of dialects rather than languages, thus restricting their
applicability. In short, it was not Stiernhielm’s interpretation of the conceptual
pair in terms of substance and accidents that was exceptional, but the fact that
he connected the Aristotelian criterion explicitly and abundantly with specific
linguistic features.
⁶ On the Scythian theory, its genesis, and its different interpretations see e.g. Metcalf (2013: 34–9);
Swiggers (1984); Villani (2003); Considine (2010); Van Hal (2010a: especially 335–401, 473–5).
⁷ For Stiernhielm’s dialect concept see e.g. already Metcalf (2013: 49–50); Coseriu (1975: 8–9); Law
(2003: 261); Burke (2004: 23); Eskhult (2020; forthcoming b). Swartling’s (1909) biography discusses
his linguistic interests passim (pp. 35–6, 70–1, 76–84).
164     
Stiernhielm’s adherence to the Aristotelian interpretation surfaces in his usage

of a variation on the ‘to differ only in dialect’ phrase, too. He argued that Aramaic,
Canaanite, Egyptian, Arabic, and Ethiopian ‘not at all differ from Hebrew except
in slight dialect’.⁸ The substantial or accidental nature of linguistic differences had
consequences for mutual intelligibility, according to Stiernhielm, who linked these
two criteria almost naturally (cf. Coseriu 1975: 9; Law 2003: 261; Chapter 8).
Speakers of distinct languages wanting to communicate needed an interpreter,
whereas speakers of related dialects were able to converse directly, even if they
might encounter occasional difficulties. This last specification suggests that
Stiernhielm regarded mutual understanding as a relative rather than an absolute
binary property. In fact, Stiernhielm admitted in an interesting 1651 letter to his
French Protestant colleague David Blondel (1591–1655) that sometimes dialectal
diversity was so strong that mutual intelligibility was impossible. He cited
Dalecarlian, the peculiar dialect of the Swedes from his native Dalarna region, as
a striking example of such a lack of mutual understanding (Stiernhielm 1937–48:
137–8).
Stiernhielm moreover viewed dialects as speech forms deviating from a point of
reference, either a common language or a shared ancestral language. The latter
seems to reveal that for Stiernhielm the ultimate foundation of the distinction was
a genealogical one, which would be in keeping with his strong language-historical
interests: a dialect was produced by, and therefore inherently posterior to, a
language. The image of a language as a tree trunk with dialects for branches was
invoked to make this idea more concrete, a metaphor of which I offer an
impression in Figure 12.2. In this context, Stiernhielm described the evolution of
a language into dialects, which in turn developed into separate languages, as an
inevitable natural process, even occurring when no foreign language was interfer-
ing. For this continuous change Stiernhielm (1671: d.4, e.1–2) posited two
main causes: distance in time and space.
Dialect B
Di
tA al
ec ec
al tC
Di
Language
Figure 12.2 Impression of Georg Stiernhielm’s tree metaphor
⁸ Stiernhielm (1671: c.3): ‘nil nisi levi dialecto, differre ab Hebræa’.

.       165
Elsewhere in his preface, Stiernhielm (1671: d.4) bracketed dialect together

with province, suggesting that he associated the concept with a delimited geo-
graphical entity. Finally, he sometimes seems to have implied that a dialect was a
faulty speech form, corrupted from a language, even though he did not seem to
present it as an inherent property of a dialect. This interpretation can, however, be
read into his usage of Latin phrases such as degenerare in dialectum, ‘to degenerate
into a dialect’.⁹
In short, Stiernhielm provided his reader with many clues about his conception
of the conceptual pair, which he centred around one main diagnostic criterion
with language-historical implications: languages mutually differ in substance,
which he principally connected to words and roots, whereas related dialects
vary only in accidental terms. His account differed significantly from the often
short and vague discussions of his sixteenth- and early seventeenth-century
predecessors.
Initially, the conceptual pair was closely tied up with the Greek language and
scholarly tradition. In the course of the seventeenth century, however, theorizing
on the conceptual pair was increasingly stripped from its Greek luggage, as I have
tried to show in Chapter 11. What was Stiernhielm’s relationship with the Greek
language and heritage? It immediately catches the eye that he barely mentioned
Ancient Greek dialectal diversity. His acquaintance with the language seems to
have been rather slight, to say the least. He limited himself to noting that Greek
could be seen as a dialect of the Scythian protolanguage on the same level as Saxon
and Gothic, both ideas for which he was relying on contemporary sources.¹⁰ In his
explanation of the conceptual pair, Stiernhielm (1671: c.3–d.1, c.4-, d.4) in
fact principally referred to the so-called Oriental languages as well as to Western
and Northern European vernacular languages such as Italian, French, German,
and his native Swedish tongue. It is striking that he did not name a single Greek
dialect, whereas he did attribute dialects to the Hebrew language, an idea for which
he was inspired by the shibboleth incident recounted in the Book of Judges. The
fact that the Ephraimites were unable to pronounce the [ʃ] sound and when asked
to say the word shibbóleth said sibbóleth instead was for Stiernhielm (1671: d.2)
a letter permutation and therefore evidence of dialectal diversity within Hebrew
(see Chapter 3, Section 3.4).
This outline may create the somewhat misleading impression that Stiernhielm
was highly original in his ideas on the conceptual pair, all the more so since he did
not refer to any sources in his two lengthy definitions. This omission must be
viewed in relation to the fact that like most early modern scholars he considered
⁹ For this phrase see the first tree diagram in Stiernhielm (1671), inserted after siglum f.4 in the
edition I consulted in April–May 2016 (Gotha, Forschungsbibliothek, Theol 4o 00022e/01).
¹⁰ See Stiernhielm (1671: e.4), where both Casaubon (1650: 139) and Junius (1665: *.3) are quoted.
Cf. also Eskhult (forthcoming b: 7.3.4) for a passing reference to the Greek dialects.
166     
the language/dialect distinction to be an evident datum; it need not be further

problematized. It is nonetheless possible to identify a probable source of inspir-
ation for Stiernhielm’s interpretation of the two concepts. His foregrounding of
the Aristotelian criterion was likely inspired by the thought of the ambulant
intellectual Christian Ravis, who expressed comparable ideas precisely in the
years 1646–50, as I have contended in Chapter 8 (Section 8.1). As a matter of
fact, Stiernhielm must have met this Berliner orientalist at the court of Queen
Christina of Sweden in the early 1650s, right around the time he was drafting the
first version of his preface on the origin of languages. He did not, however,
mention Ravis by name. Another important source for Stiernhielm was the
French classical scholar Claude de Saumaise, whose Commentary on the
Hellenistic tongue of 1643 he did quote favourably in his preface when treating
the emergence of new languages out of dialects (see Stiernhielm 1671: e.1-,
quoting Saumaise 1643a: 18–19). In this work Stiernhielm may have found
confirmation for his Aristotelian interpretation of the conceptual pair, as
Saumaise (1643a: e.g. 356, 367) repeatedly employed variations on the phrase ‘to
differ only in dialect’. Stiernhielm might have discussed the matter orally with
Saumaise, who also spent some time at Queen Christina’s court in 1650–1.
Whereas Stiernhielm’s relationship to his sources remains somewhat unclear,
there is no doubt that his views on the conceptual pair influenced the thought of
several later scholars. They were, most notably, picked up in the well-known
dissertation On Europe’s oldest language, published in 1686 in Wittenberg and
probably co-authored by Georg Caspar Kirchmaier (1635–1700) and Andreas
Jäger (c. 1660–1730; see Considine 2008b on the issue of authorship). Kirchmaier,
the supervisor, and Jäger, the presenting student, repeatedly referred to
Stiernhielm’s preface in this work. In fact, their lengthy answer to the question
‘How does a language differ from a dialect?’ was quoted almost verbatim from
Stiernhielm.¹¹
12.3 Conclusion
Georg Stiernhielm devoted considerable attention to the conceptual pair, which he

defined first and foremost in Aristotelian terms: distinct languages differ from
each other in substance, whereas related dialects do so only accidentally.
Stiernhielm was remarkable in that he tied both poles to specific linguistic
categories—language to roots and words, and dialect to so-called accidental
properties such as pronunciation and letter mutations. He noticed that this
distinction also had implications for mutual intelligibility among speakers of
¹¹ See Kirchmaier and Jäger (1686: 17): ‘Lingua quomodò à dialecto differat? exposuit in ad
Ulphilam præfat[ione] D[omi]n[us] Georg[ius] Stiernhielmus’.
.  167
distinct languages as opposed to related dialects. His language-historical pro-

gramme almost naturally led him to conceive of the relationship between language
and dialect in diachronic terms, too. More in the margin, he associated dialect
with province and with degeneration. In short, he conflated a variety of interpret-
ations in his conceptualization of the language/dialect distinction, but it is safe to
say that the Aristotelian interpretation was his most important criterion, since it
featured prominently in both of his extensive definitions of the conceptual pair.
In outlining his views, the Swedish scholar, who exceptionally defined both
dialectus and lingua, went about much more systematically and provided more
thorough and elaborate analyses than his Renaissance predecessors. The criteria
by means of which language could be distinguished from dialect were discussed
more explicitly and were linked closely to linguistic features, a connection largely
absent from earlier thought. Stiernhielm may therefore rightly be regarded as a
transitional figure. He stood at the end of the consolidation period, during
which the conceptual pair was definitively established as a metalinguistic tool
through the explicit development of its main interpretations, and heralded a
new phase in the history of the language/dialect distinction. Fully emancipated
from the Greek heritage, theorizing on the conceptual pair now became much
more systematized as well as rationalized. Before I treat both tendencies in more
general terms in Chapters 15 and 16, I will elucidate them by means of two
revealing case studies of eighteenth-century scholars with a deep concern over
the language/dialect distinction, one a productive Dutch orientalist, the other a
prominent German historian.
IV
SYSTEMATIZATION AND
R A T I O N A L I Z A T I O N , 1 6 5 0– 1800
13
Putting the conceptual pair on
the scholarly agenda
The orientalist Albert Schultens
On 22 June 1665, a German orientalist by the name of Levinus Warner ( c. 1618/9)

died childless in Istanbul. Out of gratitude Warner left his manuscript collection
to his alma mater in the Low Countries, Leiden university (see e.g. Vrolijk et al.
2012). His library contained over 900 books, predominantly in Arabic, but also
in Greek, Hebrew, and Armenian, which Warner had found during his many
years of service in Ottoman territory, first as secretary of a Dutch merchant and
later as diplomat for the Dutch Republic. The books arrived in Leiden in
different shipments between 1668 and 1674, but a lack of expertise and interest
caused them to lie forgotten in a sweat cloth for decades.
Fast forward to the summer of 1709: Albert Schultens (1686–1750; Figure 13.1),
a 22-year-old scholar from Groningen, obtained his doctorate in theology at the
academy of his hometown after studies there, in Leiden, and with the renowned
orientalist Adriaan Reland (1676–1718) in Utrecht.¹ Schultens did so with a
dissertation on a New Testament verse in the Gospel of Mark, in which he
demonstrated his excellent mastery of Latin, Greek, and Hebrew. He was, how-
ever, not a classical trilingual man, since he also mastered Arabic, a tongue by
which he was fascinated from his early years. A teenage prodigy, he matriculated
at Groningen university at the age of fourteen and professed his love for Arabic in
a dissertation held there when he was only nineteen years old. Indeed, he must
have become convinced at an early stage that Arabic was the key to arriving at a
better understanding of Hebrew.
Schultens’s enthusiasm for Arabic no doubt also explained his return to Leiden
in late 1709, intrigued as he was by the Warner treasure, which he must have seen
in that city when he studied there some years earlier. Schultens was allowed to
investigate the Warner collection in peace, which cannot but have stimulated his
interest in the Arabic tongue. Yet to his regret he had to combine this study, from
1711 onwards, with a job as pastor in Wassenaar near Leiden. To escape this duty
he accepted, in 1713, the chair of Hebrew and antiquities at the university of
¹ Biographical information on Schultens is principally drawn from the very useful but positively
biased account of Wensinck (1921).
DOI: 10.1093/oso/9780198845713.001.0001
172        
Figure 13.1 Albert Schultens, 1730

Franeker in the north of Frisia, where he further developed his idea that Arabic
could elucidate Hebrew and started to put it into practice. Scholarship on Hebrew,
Schultens believed, required a new impetus and perspective achievable by study-
ing its lexicon through the lens of Arabic, which, he complained, had thus far been
studied solely in isolation from Hebrew. This comparative approach was all the
more essential since the Hebrew text corpus was largely limited to the Old
Testament. After all, one did not study the Latin lexicon only through the works
of Cicero either. Schultens devoted most of his energies to the comparative study
of the Semitic lexicon, primarily through Hebrew and Arabic. A first major product
of his research was a volume on Hebrew etymology published in Franeker in
1724. At this stage, he still saw Arabic as a daughter language of the sacred
Hebrew language.
The initial results of Schultens’s studies impressed the Dutch scholarly com-
munity, and in 1729 he received a double offer from Leiden university, which he
gladly accepted. He became rector of the Collegium theologicum, an institute for
impecunious theology students, and, more importantly, he took up the newly
created position of interpres legati Warneriani, ‘interpreter of Warner’s bequest’.
        173
This appointment put him in a prime position to study Warner’s manuscript

collection. At the same time, Schultens’s advanced study of Hebrew and Arabic
forced him to revisit his ideas on their relationship. He now argued, more
correctly from the perspective of modern scholarly insights, that Arabic was a
sister of Hebrew rather than a daughter, both descending from a now lost
protolanguage. This new insight was, however, controversial, and his colleagues
met it with bitter reproach, as most of them still considered Hebrew to be the
uncontested primeval language of mankind (cf. Chapter 10, Section 10.2).
Schultens remained in Leiden for the rest of his life, becoming professor of
Oriental tongues in 1732 and of Hebrew antiquities in 1740. He continued to
publish on Hebrew and Arabic, even though he did not manage to have all of
his work printed. Some manuscripts of his, including his Theses on the primeval
tongue, lie unedited at Leiden University Library.² Apart from more practical
applications of his method, including editions and translations of Hebrew and
Arabic texts and a grammar of Hebrew, he also produced a second volume on
Hebrew etymology and a polemic ‘program of his principles for language
research’, both published in 1738.³ Indeed, Schultens’s new method, especially
the prominence of Arabic in it, was not appreciated by fellow theologians such
as Jacques Gousset (1635–1704) and Anton Driessen (1684–1748) who claimed
that for their discipline the study of Hebrew in itself sufficed (Eskhult 2015:
78–81). Schultens took it as his duty to defend his approach in several of his
later works, and he did so successfully. Partly thanks to the scholarly dynasty he
was able to establish at the Leiden academy through his son Jan Jacob
(1716–88) and his grandson Hendrik Albert (1749–93) after his death on 26
January 1750, his method left an indelible mark on Oriental studies in the
eighteenth century.⁴
Schultens’s profound interest in determining the precise relationship between
Hebrew and Arabic heightened his sense of the concepts with which he was doing
so. Especially in his later work, he was concerned over the metalinguistic tools he
applied to this end. One of the concepts with which he engaged most intensively
was dialect. As a matter of fact, he defined it at several occasions in different
publications, offering the modern reader a transparent window on his ideas about
the distinction between language and dialect. In this brief case-study chapter,
I address two principal questions. Firstly, how did Albert Schultens conceive of
the conceptual pair? Secondly, in what ways did his account differ from earlier
discussions? To answer these questions, it is convenient to concentrate on his two
² The Theses will, however, soon be available in an edition procured by Josef Eskhult (forthcoming a).
³ See, respectively, Schultens (1738b) and Schultens (1738c). The description of Schultens (1738c) is
Eskhult’s (2015: 74).
⁴ On Schultens’s comparative method see especially Eskhult (2015; forthcoming a). On Schultens’s
school of Oriental studies see Nat (1929: 37–103).
174        
main publications of 1738: his second volume of Hebrew origins and his Ancient
and royal way of Hebraizing.⁵
13.1 Schultens’s definitions of dialect
In both of his 1738 works, Schultens discussed at length what he considered to be a

dialect, which he did in the context of his new comparative approach to the
Hebrew tongue. He believed it necessary to elaborate in detail on the matter,
since learned men had thus far failed to understand correctly what a dialect was
(Schultens 1738b: *.2). Convinced that the Oriental tongues differed from one
another in much the same way as the Greek dialects did, he explained his
interpretation of the term as follows in the second volume of his Hebrew origins,
a work already finished by the end of 1737:
. A dialect is without doubt nothing else and cannot be anything else in the
view of all those who find pleasure in speaking accurately than a certain external
and accidental variation of one language, which does not pertain to its internal
substance, but preserves the foundation untouched and unharmed. Certainly,
when this is ruined, altered, and corrupted, it will no longer be a dialect, but a
corrupted and impure offshoot, such as that of the Italians and other peoples
who have more or less corrupted and polluted the Latin language, overthrown by
barbarian invasions from its foundations, so to speak.
. These external variations, which happen to a true dialect, can be very
conveniently referred to three classes. The first revolves around the elements of
the letters, sounds, and manner of pronouncing. The second extends itself to the
meanings of the words themselves. The third embraces syntax, patterns of
speaking, and the entire discursive course.⁶
In his other work of 1738, which outlined his research programme and post-
dated his Hebrew origins, Schultens repeated his discussion in an even more
schematized and systematized fashion, structuring it as a series of eight theses:
⁵ On Schultens’s views see also Eskhult (2015: 86–7; forthcoming a).

⁶ Schultens (1738b: 20): ‘. Dialectus nempe nihil aliud est, nec esse potest apud omnes qui
accurate loqui amant, quam Unius Linguæ variatio quædam externa & accidentalis, quæ ad internam
ejus substantiam non pertingit, sed fundamentum integrum illibatumque conservat; hoc enim
labefactato, alterato, corrupto, non jam amplius Dialectus erit, sed corrupta & adulterina soboles,
qualis est Italorum aliorumque populorum, qui Latinam Linguam invasionibus Barbarorum tanquam a
fundamentis suis convulsam plus minusve corruperunt atque adulterarunt.
. Hæ variationes externæ, quæ in veram Dialectum cadunt, commodissime ad tres classes
possunt referri. Prima versatur circa elementa literarum, sonos, ac pronunciandi modum. Altera ad
significationes ipsas verborum sese extendit. Tertia complectitur constructionem, loquendi formas,
totumque orationis ambitum’.
. ’    175
. A dialect is a variation of one language which is only external and accidental

and does not touch its internal substance, let alone that it taints and
corrupts it.
. If the root and marrow are tainted, they no longer have to be called dialects,
but degenerate offshoots which have flowed away to an entirely other
nature.
. Whoever would call the Latin language, even though it is connected to the
Greek by a very strong bond, a dialect of the same language and would add
this language as a fifth to Attic, Ionic, Doric, and Aeolic, would gain praise
neither of skill nor of judgement.
. The same mistake is made by those who are accustomed to call Italian,
Spanish, French, and English speech dialects of the Latin language.
. One commits an even bigger blunder, when proposing nearly the same
relationship of the Italian, Spanish, French, and English speech vis-à-vis the
Latin mother, as of the Chaldean, Syriac, and Arabic dialect vis-à-vis
Hebrew.
. Finally, one makes the greatest crash, when one, together with the famous
Gousset and Driessen recently, thence, as it were from an unconquerable
stronghold, assails and attacks the use of the Oriental dialects in elucidating
Hebrew.
. The affinity and indeed blood-relationship of the Chaldean, Syriac, and
Arabic dialects with Hebrew is equally close as that of Ionic, Doric, and
Aeolic with Attic.
. Assuredly, whoever nevertheless in this way denies that the Hebrew lan-
guage is to be clarified by means of the dialects, denies the principles which
are beyond all doubt conveyed, and these they could not or would not want
to call into doubt anywhere else.⁷
⁷ Schultens (1738c: 109):

. Dialectus est unius Linguæ variatio tantum externa, & accidentalis, quæ internam ejus
substantiam non attingit; nedum ut eam vitiet, atque corrumpat.
. Si Radix & Medulla vitietur, non amplius Dialecti appellari debent, sed degeneres propa-
gines, quæ in aliam plane indolem defluxerunt.
. Qui Latinam Linguam, triplici licet vinculo Græcæ innexam, appellaret ejusdem
Dialectum, atque Atticæ, Ionicæ, Doricæ, Æolicæ, accenseret quintam hancce, nec
peritiæ, nec judicii laudem auferret.
. Idem peccant, qui Italum, Hispanicum, Gallicum, Anglicum Sermonem, Linguæ Latinæ
Dialectos appellitant.
. Gravius adhuc offendunt, quum Italici, Hispanici, Gallici, Anglici Sermonis eandem fere
statuunt relationem erga Matrem Latinam, quam Chaldaicæ, Syriacæ, & Arabicæ Dialecti
erga Hebræam.
. Gravissime denique impingunt, quando cum Cl. Gussetio, & Driessenio nuper, inde,
tanquam ex Arce invicta, usum Dialectorum Orientalium in Hebræa illustranda impetunt,
atque oppugnant.
. Adfinitas, atque adeo consanguinitas Dialectorum Chaldaicæ, Syriacæ, Arabicæ, cum
Hebræa, æque arcta est ac Jonicæ, Doricæ, Æolicæ, cum Attica.
. Qui vel sic tamen negat Linguam Hebræam ex Dialectis illustrandam, Principia negat supra
omnem aleam subvecta: & quæ nuspiam alibi posset velletve in dubium revocare’.
176        
The definitions reveal that Schultens was fairly traditional in interpreting the
distinction between language and dialect. His views remind of Georg Stiernhielm’s
to a certain extent, in that the contrast centred around the Aristotelian criterion
(see Chapters 8 and 12). As a matter of fact, Schultens was familiar with the
Swede’s work, which he quoted in his manuscript Theses on the primeval tongue
(Eskhult forthcoming a: §). Even though the Aristotelian interpretation took
centre stage in his definitions, Schultens (1738b: 104; 1748: ) also seems to
have associated dialect with limited geographical coverage and used the conceptual
pair in language-historical terms in his work. Before him, Stiernhielm, too, had
foregrounded the meaning in terms of substance and accidents in his definition,
although also assuming that dialect had limited geographical coverage and genea-
logically derived from a language.
Schultens displays several notable differences vis-à-vis Stiernhielm, however.
Let me highlight the three most important ones for our purposes. Firstly,
Schultens seems to have associated language more closely with analogy, a feature
dialects lacked, even though they could help to uncover the original analogy of a
language. He did so in the long preface to one of the practical applications of his
method, his edition of the Hebrew text of the Book of Proverbs with a Latin
translation and extensive commentary (Schultens 1748: –). Secondly,
Schultens did not regard mutual intelligibility as an intrinsic characteristic of
related dialects. Instead, he suggested that the more time passed, the more it
decreased, that is to say, when a language gradually and insensibly—‘sensim sine
sensu’, one of his favourite phrases in Latin—developed into several dialects.⁸
A third difference with Stiernhielm was that Schultens seems to have been well-
informed about the Greek dialects, which he compared ad nauseam with Oriental
diversity.⁹ He paralleled the four Greek dialects Attic, Ionic, Doric, and Aeolic to
Hebrew, Aramaic, Syriac, and Arabic. This Greek–Oriental linking indicates that
what I have dubbed the de-Hellenization of the conceptual pair was a gradual and
uneven process. Even though Stiernhielm had already proposed a definition
without referring to the Greek background, Schultens, a trained Hellenist, still
frequently resorted to it (cf. Chapters 11–12).
13.2 Language, dialect, and degenerate offshoot
If it is not in his interpretation of the language/dialect pair that Schultens was

innovating, where was he? I argue that he was responsible for two major
⁸ See e.g. Schultens (1739: 235; 1748: ; in Eskhult forthcoming a: §, §, §); Eskhult
(2015: 87).
⁹ See e.g. Schultens (1738a [1729]: 36; 1732: 4; 1737: 5; 1738b: 21, 104; 1738c: 106–8; 1739: 187, 189,
190–1, 222, 234; 1748: , ; in Eskhult forthcoming a: §, §, §§–).
See also Nat (1929: 39).
. , ,    177
innovations with consequences for the history of the two concepts. Firstly,
Schultens’s definitions were part of an attempt at developing a linguistic method
to determine whether a variety was a dialect or a degenerate offshoot of a
language, taken here in the language-historical sense of ‘common ancestor’
(Chapter 10). In other words, Schultens operated with a tripartite hierarchy with
implications for linguistic chronology (Figure 13.2) rather than with a two-way
conceptual distinction.
Language
(e.g. Greek, Latin, primeval language)
Dialect
Corruption
Time
(e.g. Aeolic, Arabic)
Degenerate offshoot
(e.g. French)
Figure 13.2 Impression of Albert Schultens’s tripartite conceptual hierarchy
Schultens differentiated a dialect from a degenerate offshoot in terms of purity.

The latter was corrupted, whereas the former preserved the foundation of the
ancestral language intact. On the one hand, his positive conception of dialect is
likely to relate to the long-standing literary traditions of the speech forms he
termed dialects: varieties of the Greek and Oriental language families. On the other
hand and more importantly so, the distinction was informed by Schultens’s
research agenda. After all, the argument that the Oriental tongues were uncor-
rupted dialects of a now lost primeval language served to justify the fundamental
principle of his language studies, a comparative approach to etymology, which in
Schultens’s eyes made sense only if the comparison was based on evidence from
uncorrupted varieties. In this context, Schultens repeatedly referred to a linguistic
situation he claimed did not exemplify dialectal variation, the relationship of the
Romance tongues to the Latin language, which he analysed in terms of corruption
in a typically early modern manner (see e.g. Neis 2009). His argument implied that
the Romance tongues could not be unproblematically used for the study of Latin
etymology.
Remarkably, Schultens included English among the Romance tongues, which
might be explained by the fact that this language has many French and Latin
loanwords as well as by the fact that Schultens’s knowledge of it must have been
rather limited. Schultens found it apparently important as well to point out that
Latin was not a dialect of Greek. Acknowledging their close relationship, he
178        
refused to follow the age-old tradition going back to antiquity of so-called

Aeolism, according to which Latin derived from Greek through its Aeolic dialect
(Stevens 2006–7).
13.3 Classes of linguistic variation
A second major innovation transpires from the specifications Schultens added to

his definition of dialect in his Ancient and royal way of Hebraizing, which elaborated
on an earlier brief remark made in his second volume of Hebrew origins:
Certainly those external variations, which produce a true dialect, happen either to
the elements or manner of pronouncing, or to the meanings of words, or to the
entire construction, whence they can all be referred very conveniently to three
classes.
. In the first class, diversities which play around the bark exhibit themselves,
1. by means of a letter change;
2. by means of a vowel change;
3. by means of a slight transposition of both;
4. by means of a contraction of form;
5. by means of a lengthening of the same;
6. by means of a shifting of the accent.
. The second class, which pertains to the leaves and foliage, produces even more
variation.
1. In some domestic words, which each people is found either to have received
as proper to itself or to have coined for itself. A rare kind.
2. Very frequently, there is variation in secondary meanings, which are
introduced through metaphors and other tropes.
3. Hence, such a great distinction often exists, that there seems to be a blatant
contradiction between the significations of words which in their origin
nevertheless agree very pleasingly.
4. One dialect preserves that origin better than the other. Thus, just as in one
dialect there are very few [primary meanings], the same can be found very
copiously in the other.
. The third class of variations is constituted by:
1. syntax and the fashion of construction, which is often very discordant with,
and different from, the innermost agreement of meanings;
2. formulas which inevitably arise in each people out of its domestic rites and
affairs;
3. the entire phraseology, in such a manner adjusted to the genius of each
people that just as an Attic writer is immediately distinguished from an
.     179
Ionic, a Syrian author can be likewise immediately distinguished from an

Arab, and both of them from a Hebrew author.¹⁰
To paraphrase, related dialects exhibited only superficial or external variations,

which concerned firstly letters and their pronunciation—the bark of a tree—
secondly the meaning of words—the foliage of a tree—and rarely the words
themselves, and thirdly syntax, which Schultens did not equate with a tree part.
The first class seems to be a variation on the age-old permutation of letters
framework, comprising pronunciation and covering also such features as accent
and prosody (Chapter 4, Section 4.4). For the second class, the variation of
meanings of words across related dialects, Schultens may have drawn on early
modern handbooks of Ancient Greek and its dialects, in which this type of
semantic variation was frequently mentioned, even if no obvious source can be
singled out.¹¹ What is more, he seems to have been one of the first scholars to
include it in a systematic treatment of the levels of dialectal variation outside such
Greek-centred handbooks. To further refine this class of variation, he made a
distinction between primary and secondary meanings, which varied across
dialects. He assumed that some dialects preserved the original meanings better
than others, and that individual dialects tended to introduce different secondary
meanings. The third class involved syntax, which Schultens mainly identified with
phraseological and idiomatic properties, features he linked to the customs and
¹⁰ Schultens (1738c: 110–11): ‘Nempe variationes illæ externæ, quæ veram Dialectum efficiunt,
incidunt vel in Elementa & pronunciandi modum; vel in significationes verborum; vel in constructionem
universam: unde ad tres classes cuncta commodissime revocari possunt.
. In prima classe offerunt se diversitates circa Corticem ludentes,
1. Ex immutatione literæ.
2. Ex vocali permutata
3. Ex levicula trajectione utriusque.
4. Ex contractione formæ.
5. Ex productione ejusdem.
6. Ex Accentu aliter & aliter posito.
. Altera classis ad folia & frondes pertingens, adhuc plus varietatis parit.
1. In vocibus quibusdam domesticis, quas quæque gens sibi vel proprias adscivisse, vel novasse,
reperitur. Genus rarum.
2. Frequentissime variatum in significationibus secundariis, quæ per Metaphoras, aliosque
Tropos, invectæ.
3. Tantum hinc discrimen sæpe existit, ut directa fronte pugnare videantur potestates verborum,
quæ in Origine tamen amicissime conspirant.
4. Istius Originis alia Dialectus præ alia servantior: sic ut in una paucissimæ, in altera
copiosissimæ, eædem inveniantur.
. Tertiam classem varietatum constituunt:
1. Syntaxis, & ratio constructionis sæpe cum intima significationum concordia valde discors, &
dissonans.
2. Formulæ ex ritibus rebusque domesticis in quaque gente oriundæ.
3. Phraseologia universa, cujusque gentis genio ita attemperata, ut quemadmodum Atticus
scriptor ob [sic] Ione primo adspectu discernitur, sic Syrus ab Arabe, & uterque ab Hebræo
Auctore statim discriminari queat’.
¹¹ Possible sources include Zwinger (1605: .2) and Ursin (1691: 496).
180        
genius of the people speaking the dialect. Schultens did not make explicit how
distinct languages varied exactly, which is in line with the broader early modern
tendency of leaving the overarching concept of language usually undefined. In
sum, Schultens was rather exceptional in devoting extensive and systematic
attention to the three ‘classes’ of variation in which related dialects allegedly
differed from one another.
13.4 Schultens’s legacy
As I have mentioned before, Schultens was able to establish a scholarly dynasty at

Leiden university, with his offspring continuing his philological programme.
However, his influence reached far beyond his own family, both within and
outside the Dutch Republic. Schultens was the teacher of numerous students
who managed to build successful careers as orientalists. Was he also able to instil
into them his ideas about dialect and how it related to language and degenerate
offshoot? It most certainly seems so. Two convincing pieces of evidence can be
cited: his teaching and the writings of his pupils.
Schultens must have professed his views on what constituted a dialect not only
in his writings, both published and manuscript, but also in his Hebrew courses at
Leiden university. A unique manuscript preserved at the library of this institution
contains lecture notes ‘on Hebrew origins and other topics’ taken by the 23-year-
old Lodewijk Caspar Valckenaer (1715–85), a Frisian student of Schultens, who
later made a name for himself as a Hellenist.¹² On the fifth folium of this
manuscript, Valckenaer noted some of Schultens’s ideas about the interpretation
of the term dialectus. These notes, taken right before rushing to Friesland on 6 July
1738, as Valckenaer informs us, are very much in line with the views Schultens
expressed elsewhere in his published work, especially, nota bene, in his two
monographs published in the same year in which Valckenaer wrote down his
notes. These student notes suggest that Schultens did not limit himself to discussing
the topic in his publications, but he also actively taught it to the audience of his
Hebrew courses at Leiden university.
Also in their published work several of Schultens’s pupils prove to have been
influenced by their teacher. A number of his students, including the Hellenists
Tiberius Hemsterhuis (1685–1766) and Lodewijk Caspar Valckenaer, as well as
the orientalist and missionary preacher Jan Willem Kals, discussed the concept of
dialect and related issues in their work, albeit to various degrees of dependency
on their master’s ideas. Whereas Schultens’s influence on the ideas of the
Hellenists Hemsterhuis and Valckenaer was at most limited, the orientalist
¹² See Leiden, University Library,  469., fols. 1–23, ‘de originibus Hebraeis et aliis’.
.  181
Kals’s conception of dialect as opposed to language was very much indebted to

Schultens’s views (van der Linde 1987: 98). Defining the Latin terms lingua and
dialectus in the spirit of his teacher, Kals (1752: 9–10) claimed that a correct
definition of them was only of secondary importance. Instead, he mainly
attempted to prove, very much unlike his former professor, that Hebrew in fact
was the primeval language (Kals 1752: 27, 61; van der Linde 1987: 101). He also
criticized Schultens for propagating two contradictory views, failing to see that
Schultens had actually distanced himself from his initial idea that Hebrew was the
primeval language (Kals 1752: 62). Faithful to Schultens’s method, Kals (1752: 67)
nevertheless admitted that Hebrew had only been partly preserved, stressing that
its dialects were necessary to recover its lost part. He also agreed with his master
that language change occurred gradually and insensibly, even taking over his
favourite phrase verbatim (Kals 1752: 9–11). Kals’s work, published in Oxford,
extended Schultens’s influence far beyond the borders of the Dutch Republic.
The list of Schultens’s students influenced by his idea of what constituted a
dialect can be enlarged further still. For instance, his Swiss theology student
Antoine-Noé de Polier de Bottens (1713–83) authored a dissertation, orally
defended in Leiden under Schultens’s supervision in 1739, on the purity of
Arabic in comparison with Hebrew, which he both saw as descending from the
primeval tongue. Polier consciously opted to speak of the Arabic dialect and not
language, a contrast requiring explanation to his mind. The definition of dialectus
he provided was very obviously influenced by that of his esteemed teacher (Polier
de Bottens 1739: 4).
The evidence presented here suggests that Schultens actively encouraged his
students to reflect on the concept of dialect. He might have discussed his ideas
with travelling scholars, too. It is, for instance, not unlikely that he met the
Danzig-born orientalist Benjamin Groddeck (1720–76) passing through the Low
Countries in the 1740s. At any rate, Groddeck was the supervisor of a dissertation
defended in 1747 in Wittenberg, in which the nature of dialects was discussed with
special reference to Hebrew and Arabic, and for which Schultens’s conception of
the term and his ideas on the Oriental tongues constituted without doubt the
primary source of inspiration (compare Schultens 1738b: 20–1 with Groddeck and
Treuge 1747: –).
13.5 Conclusion
The Dutch orientalist Albert Schultens took exceptional care to procure his reader
with a systematized account of the semantics of the language/dialect pair.
Schultens did so because he was dissatisfied with existing definitions and usages
of the term dialectus. He went beyond a simple binary contrast, however, trans-
forming the conceptual distinction into a tripartite hierarchy consisting of
182        
language, dialect, and degenerate offshoot. His interpretation of these concepts

revolved around the Aristotelian criterion, with a clear language-historical dimen-
sion, and was, as in Stiernhielm’s work, closely linked to linguistic features,
‘classes’ of variation in Schultens’s own words. His tendency to systematize
theorizing on the issue is not only apparent from the content of his works but
also emerges from the way in which he presented his ideas typographically.
Indeed, his discussion was visualized in a highly schematized fashion, as one
can gather especially from his Ancient and royal way of Hebraizing (see Schultens
1738c: 109–10 in particular).
Schultens can deservedly be regarded as a scholar who put theorizing on dialect
on the scholarly agenda, even though it remained a matter of secondary import-
ance only, as the writings of his pupil Kals suggest. The distinction between
language and dialect was indeed first and foremost a tool for early modern
scholars, who usually did not problematize it. Its auxiliary status is abundantly
clear from the work of a late eighteenth-century historian, who tried so hard to
offer an objective interpretation of the conceptual pair that he made its arbitrariness
painfully obvious.
14
Lexicostatistics avant la lettre
The historian Johann Christoph Gatterer
and the conceptual pair
The protagonists in this book have thus far been primarily philologists fascinated
with ancient texts and languages, usually Greek and sometimes Semitic. In the
eighteenth century, the linguistic horizon broadened still further. Scholars
developed an increasing interest in the dialects of other, mainly vernacular
tongues, especially in Scandinavia and German-speaking lands but also in
Britain, France, and Italy. There, high-quality grammatical and lexicographical
descriptions of dialects began to appear. One of the best-known is no doubt the
lexicon of Swedish dialects by Johan Ihre (1707–80), published in 1766. This
new interest was fuelled partly by curiosity, partly by intensifying antiquarian
and historiographical activities. Some intellectuals went to great lengths to
uncover ethnic history and looked at all possible types of evidence in their
pursuit. A case in point is the Bavarian historian Johann Christoph Gatterer
(1727–99; Figure 14.1), professor of history at Göttingen university for the
largest part of his career (1759–99) and from December 1766 onwards also
director of the Göttingen Historical Institute (see Dohna 1964 for biographical
data). Gatterer owed this position to the renown of his informal historical
seminar, in which he propagated the writing of comprehensive history. This
universal historiography should be soundly founded on the critical study of
different kinds of sources, Gatterer believed, and not only on the work of
predecessors, as was often the case in the early modern period. He stressed
the importance of ancillary sciences such as chronology, heraldry, and geog-
raphy, thus paving the way for modern historical scholarship (Grafton 2007:
189–91; Iggers 2011: xix; Gierl 2013: 285). One of the types of evidence Gatterer
valued highly was linguistic in nature. More specifically, he contended that a
critical study of languages and linguistic kinship helped scholars compose an
encompassing history of past societies.
DOI: 10.1093/oso/9780198845713.001.0001
184    
Figure 14.1 Johann Christoph Gatterer

14.1 Gatterer’s Vorrede
In March 1770, Gatterer held a discourse at the Göttingen academy in which he

emphasized the utility of languages for writing ethnic history.¹ His lecture was
published shortly thereafter under the German title Vorrede, von der historischen
Benutzung der Sprachen, or Preface on the historical usage of languages in English.
The text was dated 6 April 1770 and prefixed to the thirty-fourth volume of the
monumental General world history (Allgemeine Welthistorie), a titanic historio-
graphical project in which Gatterer was closely involved. He later integrated
sections of his Preface into his Introduction to synchronistic universal history of
1771 as part of a section on human languages, of which he provided an extensive
overview per continent (Gatterer 1771: 101–64, esp. 105–11). Gatterer intended
his Preface to be a foundational text. Clearly, he thought very highly of language as
a historiographical source, in all of its forms:
¹ Cf. Muller (1984a: 41). For Gatterer’s language-related views see Brekle et al. (1992–2005:
.216–18).
.  ’   185
There is, as is known, a large amount of languages. They are all worthy of a
philosopher’s investigation, also the languages of savages, man-eaters. Philosophi-
zing on languages is, in a sense, nothing else than philosophizing on reason itself.²
Gatterer (1770: 5) believed in a close connection between language and people,

considering the former ‘the reflection’ (die Abspiegelung) of the latter. For this
reason, language comparison had great historiographical utility. Not only did a
competence in many languages make one a better historian, but there was also a
key historiographical axiom regarding languages and peoples ‘which is not as
known as it deserves to be’: ‘Peoples speaking one single or closely related
languages belong to one and the very same ethnic stock’.³ This axiom was of
especial importance when one tried to reveal the history of prehistoric and
mythical times. Gatterer viewed himself as elaborating on Gottfried Wilhelm
Leibniz’s (1646–1716) views in the latter’s Brief account of reflections on the origins
of peoples, primarily drawn from the evidence of languages (Leibniz 1710). Leibniz
did not have the necessary instruments to state something substantial on this
matter, even though he should be praised for trying and for inciting later scholars
to imitate him (Gatterer 1770: 6; see also Trabant 2006: 190).
After his introduction, Gatterer briefly commented on specific languages and
language families with reference to earlier scholarship. In some cases, the com-
parison of languages had already occurred and the kinship of peoples had been
demonstrated, such as among the so-called Oriental tongues. Gatterer stipulated
the existence of an Oriental people on the basis of the close linguistic kinship
between Hebrew, Aramaic, Syriac, Arabic, and Ethiopian. In other cases, linguistic
comparison was still required, for instance in order to determine the precise
relationship between Romans and Greeks. Still, in both cases, definitive answers
could only be found when adequate instruments were available, by which he
meant comparative grammars and dictionaries. The comparative tables designed
by Gatterer, which he mentioned in his Preface, but which were unfortunately not
printed together with it, were intended to illustrate the historiographical use of
such linguistic tools. In this part of his Preface, Gatterer also resorted to the
concept of dialect to describe certain linguistic entities and relationships. For
instance, he stressed the conservative nature of German dialects vis-à-vis the
book language (Büchersprache), which was as mutable as fashion in clothing.
Gatterer even hyperbolically stated that one could hear Otfrid and Williram,
² Gatterer (1770: 4): ‘Es gibt bekannter massen eine Menge Sprachen: sie sind alle der Untersuchung
des Philosophen würdig: auch die Sprachen der Wilden, der Menschenfresser. Ueber Sprachen
philosophiren heist gewisser massen nichts anders, als über die Vernunft selbst philosophiren’.
³ Gatterer (1770: 6): ‘Ich sehe hier auf eine Art von historischer Benutzung der Sprachen, die nicht
so bekannt ist, als sie es zu seyn verdient. Völker, die einerley oder sehr verwandte Sprachen reden,
gehören zu einem und eben demselben Völkerstamme’.
186    
two medieval German authors, speaking among the Swabians, Franks, Bavarians,
and Austrians of his day, which he claimed to know from experience. He also
interpreted a number of what are now called the Baltic languages in terms of
dialect, just as he did with the Germanic tongues and some Celtic varieties
(Gatterer 1770: 5–7). Gatterer then moved on to more general considerations
touching on his intriguing method of language comparison and the concepts with
which he operated.
14.2 Characteristic words: Gatterer and basic vocabulary
As he wanted to put historiography on a par with natural sciences in terms of

objectivity, Gatterer (1770: 13) spelled out in detail how he believed accurate
language comparison should occur. He observed that ‘in language investigations
and also elsewhere, even in common speech, the expressions “cognate languages”,
“dialects of a language”, “identical words”, and “different words” occur’. He asked
himself in this context:
What, then, are called languages? What are called dialects? How far does the
kinship of languages need to go to be able to rightly say that these or those
languages are not merely cognate languages but only dialects? Since I have not
found anywhere a thorough philosophical discussion of this topic, I thus want to
venture some thoughts on it myself and present them to experts for further
investigation.⁴
In other words, Gatterer’s discussion was triggered by a discontent with previous,

superficial treatments of the issue. He proceeded by analysing the linguistic entity
of the word. It had a meaning, since it referred to a certain concept (Begriff), and a
form, materialized by letters (Gatterer 1770: 13–14). These two aspects were the
touchstones one ought to employ in order to determine whether a word present in
two or more languages was in fact the same or different. One should, however,
proceed methodically. Technical, religious, and onomatopoetic words should be
avoided, suggested Gatterer, thus proposing an exclusion criterion. Instead, one
should resort to what Gatterer called ‘characteristic words’ (characteristische
Wörter), the basic vocabulary of a language, to put it with a phrase from modern
⁴ Gatterer (1770: 13): ‘In Sprachuntersuchungen, und sonst auch, selbst im gemeinen Reden,
kommen immer die Ausdrücke vor: verwandte Sprachen, Dialecte einer Sprache, einerley Wörter,
verschiedene Wörter. Was heissen nun verwandte Sprachen? Was heissen Dialecte? Wie weit muß die
Verwandtschaft der Sprachen gehen, bis man mit Rechte sagen kan, diese oder jene Sprachen sind nicht
blos verwandte Sprachen, sondern sie sind nur Dialecte? Weil ich nirgends etwas gründliches hierüber
philosophirt finde, so will ich selbst hierin einige Gedanken wagen, und Kennern zu weiterer Prüfung
vorlegen’.
.       187
linguistics. This group of words, Gatterer added, included five categories accord-
ing to the consensus of all language philosophers: numerals; pronouns; the verb ‘to
be’; the most essential words denoting body parts, indispensable pieces of clothing,
tools, blood-relationships such as ‘father’, and things one can normally see on a
daily basis like ‘the sun’; and ‘root words’ (Stammwörter). These characteristic
words were about 300 in number (Gatterer 1770: 15). Gatterer was not the first
early modern scholar to work with the concept of basic vocabulary, which has its
roots in the seventeenth century, but he was highly specific about it and its
methodological use.⁵
The identity of words either was visible and therefore immediately clear or
needed to be revealed by means of a principle Gatterer called Reduction. By means
of this ‘reduction’, words seemingly different in terms of either meaning or form
or in both respects were proved to have an identical origin. In case of semantic
divergences, the reduction could be done by showing that the meaning of the
words in question derived from one basic idea or agreed by a certain kind of
analogy. When there were formal differences, reduction entailed showing that
letters had been transposed, added, or removed according to certain rules intro-
duced into the language, or were simply pronounced differently. Gatterer thus still
relied on the traditional framework of letter permutations, even though he
specified that these must be in agreement with certain language-specific rules.
He concluded this train of thought by exemplifying his distinction between visible
lexical identity and lexical identity in need of reduction (Gatterer 1770: 14–15).
14.3 Determining the degree of linguistic kinship:

Gatterer’s lexicostatistic framework
The discussion of the ‘characteristic words’ and the principle of ‘reduction’

constituted the fundament for Gatterer’s analysis of the different degrees to
which speech forms could display kinship:
Starting from this assumption, I believe to be able to provide some rules by

means of which one can determine the distinction between languages and
dialects, and between related and unrelated languages, and at the same time
the degree of kinship.
1. Languages in which only a few characteristic words are the same—whether
this identity is visible or reduced—are unrelated languages and even fewer
dialects. These few identical words merely indicate that the human race once
⁵ On the early modern concept of basic vocabulary and its history see Muller (1984a: especially 41–2
for Gatterer’s views; 1986: 17–19).
188    
spoke only one language. In some words, the identity can also be merely
accidental.
2. Languages in which either half or almost half or in any case up to one-third of
the characteristic words are identical can be regarded as related languages.
One immediately sees how the degrees of kinship can increase and decrease.
3. Languages in which more than half of the characteristic words are identical
can be regarded as dialects of one language.
4. Languages in which two-thirds or more than two-thirds of the characteristic
words are identical can be called closely related dialects and their kinship
increases or decreases, 1) depending on whether the number of characteristic
words is over or under two-thirds and 2) depending on whether the identity is
more or less visible.
The more visible the identity of the characteristic words is in terms of meaning
and letter form, the less forced the reduction is, and the more rarely the reduction
needs to happen in order to prove the identity, the surer one can be that languages
in which all this can be retrieved belong together as dialects; and the more it
happens, the more related the dialects themselves are.⁶
The word Dialect, explained Gatterer, was taken from Sprachphilosophie, ‘lan-
guage philosophy’. Sprachphilosophie referred in this context to general reflections
on language, which in the eighteenth century could comprise historical-
comparative approaches, too (cf. Van Hal 2015: 57–8). Gatterer (1770: 6) used
this terminology from language philosophy to elaborate his own linguistic ideas,
⁶ Gatterer (1770: 15): ‘Dieß vorausgesetzt, glaube ich im Stande zu seyn, einige Regeln zu geben,
wodurch man den Unterschied zwischen Sprachen und Dialecten, und zwischen verwandten und nicht
verwandten Sprachen, und zugleich die Grade der Verwandtschaft bestimmen kan.
1. Sprachen, in denen nur wenige characteristische Wörter eine Identität, sie sey nun sichtbar,
oder reducirt, haben, sind nicht verwandte Sprachen, noch weniger Dialecte. Diese wenigen
identischen Wörter zeigen nur an, daß einmal das menschliche Geschlecht nur Eine Sprache
geredet hat: in einigen kan auch die Identität blos zufällig seyn.
2. Sprachen, in denen die characteristischen Wörter entweder zur Hälfte, oder nahe gegen die
Hälfte, oder allenfalls bis zum dritten Theile identisch sind, können für verwandte Sprachen
gehalten werden. Man sieht zugleich, wie die Grade der Verwandtschaft steigen und fallen
können.
3. Sprachen, in denen die characteristischen Wörter über die Hälfte identisch sind, können für
Dialecte Einer Sprache gehalten werden.
4. Sprachen, in denen zwey Drittheile, oder mehr als zwey Drittheile der characteristischen
Wörter identisch sind, können nahe verwandte Dialecte heissen, und ihre Verwandtschaft
steigt oder fällt, 1) je nachdem die Menge der characteristischen Wörter über oder unter zwey
Drittheilen ist, und 2) je nachdem die Identität mehr oder weniger sichtbar ist.
Je sichtbarer die Identität der characteristischen Wörter in Bedeutung und buchstäblicher Gestalt ist;
je ungezwungener die Reduction ist, und je seltener die Reduction geschehen darf, um die Identität
darzuthun: desto gewisser kan man seyn, daß Sprachen, in denen alles dieses anzutreffen ist, als
Dialecte zusammengehören, und je mehr es statt findet, desto verwandter sind selbst die Dialecte’.
On Gatterer’s definitions see also Muller (1984b: 394).
.       189
which he formulated as one approach within a more encompassing historiograph-

ical programme.
It is surprising that Gatterer, even though his discussion was triggered by
historiographical concerns, offered a mainly synchronic, ahistorical analysis of
linguistic interrelationships. A consequence of this focus was that he did not seem
to consider a dialect to be posterior to a language like many of his predecessors
and contemporaries. The main criterion on which Gatterer relied to distinguish
between language and dialect was a linguistic one, and more specifically the degree
to which the ‘characteristic words’ of a speech form, which Gatterer generically
called Sprache, differed from those of other varieties. The historian Gatterer relied
more narrowly on lexical data than the orientalist Albert Schultens and put the
framework of letter permutations to use as an etymological principle rather than
as a class of variation, as Schultens had done.
The quadripartite division into unrelated languages, related languages, dialects,
and closely related dialects indicates that Gatterer was looking for objective,
linguistic, and quantifiable criteria to determine the degree of linguistic kinship
between two speech forms, with a remarkably low language/dialect threshold at
50%. In modern lexicostatistic research, this cut-off point is usually much higher
(cf. Chapter 20). Gatterer’s distinctions still remained arbitrary, however. As a
matter of fact, he was simply dividing the continuum from ‘no kinship’ to
‘identity’ into four arbitrary categories, which I have visualized in Table 14.1.
There was, in other words, a certain tension between the need for laying down
conceptual categories in a quantifiable way and the implicit recognition of their
relativity and arbitrariness. Gatterer’s pioneering attempt at quantifying linguistic
distance by lexical means fitted in with the ambitious programme of the Göttingen
school of history and its members, who frequently resorted to statistics in order to
analyse historical data (Cheng 2008: 130). Unfortunately, Gatterer does not seem
to have actively and systematically applied in his published work the embryonic
lexicostatistic method he outlined, but a closer study of his extensive oeuvre might
reveal whether it shines through in his preserved scholarly output.
Although trained in classical as well as Oriental languages (Dohna 1964: 89;
Debus 1998: 10), Gatterer refrained from referring to the Greek dialects in his
Table 14.1 Johann Christoph Gatterer’s four categories on the kinship continuum
no kinship identity
0% 33,3% 50% 66,6% 100%
unrelated languages related dialects closely related dialects

languages
190    
conceptual and methodological outline. As is apparent from his other work, he

knew that the extinct ancient language had four dialects: Aeolic, Doric, Ionic, and
Attic (Gatterer 1771: 119). He moreover tried to assess the relationship of this
language to others and considered it closely cognate to Latin, Germanic, Celtic,
and Slavic (Gatterer 1770: 7). It is nonetheless clear that, unlike Schultens but very
much in line with the general tendency towards emancipation from the Greek
heritage, Gatterer developed his thoughts independently from the Greek context.
This autonomy is not very surprising. Gatterer in fact pointed out that he was
elaborating a new conceptual framework, motivated by a lack of satisfaction with
existing discussions of the matter. His lamentation that there were barely any
dialect lexicons and grammars (Gatterer 1770: 9) should no doubt be seen in
connection with it and may betray a lack of acquaintance with the early modern
dialectological tradition in the philological study of Ancient Greek. Other scholars
in Gatterer’s Göttingen circle who engaged in language study likewise refrained
from referring to the Greek language in discussing the dialects of other tongues.
For instance, the orientalist Johann Gottfried Eichhorn (1752–1827) was an
alumnus of the Göttingen academy and member of its school of history, whose
mainly historiographical dissertation on Arabic dialects made no reference what-
soever to the Greek language (Eichhorn 1779).
14.4 The historian Gatterer and the grammarian Adelung
Gatterer’s ideas on linguistic kinship left their mark on the thought of the
influential German grammarian Johann Christoph Adelung (1723–1806). In his
1781 German grammar, Adelung suggested a threefold rather than quadripartite
conceptual division, splitting the kinship continuum into different languages,
related languages, and dialects. In doing so, Adelung partly adopted the terminology
of Gatterer, but he linked it to other linguistic features than the historian had done:
If two languages mutually agree in their root words and their flexion and
derivation syllables entirely, i.e. with a few exceptions, and the difference merely
exists in vowels and cognate consonants, then they are only dialects of each other.
If the deviation also covers consonants other than cognate ones and there are
noticeable differences in derivations and flexions, then they are related languages,
a relationship which has, in its turn, its manifold degrees. Entirely different kinds
of derivation and flexion and greater or smaller difference in roots and their
meaning result in more or less different languages.⁷
⁷ Adelung (1781: 71): ‘Wenn zwey Sprachen in ihren Wurzelwörtern, Biegungs- und
Ableitungssylben im Ganzen, d. i. biß auf einzele Ausnahmen, mit einander übereinstimmen, und
der Unterschied bloß in den Hülfslauten und verwandten Hauptlauten bestehet, so sind sie bloße
.  191
Instead of operating solely with Gatterer’s characteristic words, Adelung opted to

include in his analysis not only ‘root words’ but also other linguistic features. He
considered morphological differences in particular to be an important factor in
the constitution of distinct languages, whereas many previous scholars had
included flexional variation primarily under the traditional framework of letter
changes on the level of dialect.
In his Detailed doctrine of the German language, published in 1782, Adelung
repeated the statement from his German grammar quoted above. Here, however,
it was preceded by an extensive quote from Gatterer’s work. More specifically, he
cited Gatterer’s (1771: 106–11) entire explanation of his method for determining
linguistic kinship (Adelung 1782: 239–44). The way Adelung introduced the quote
suggests that he admired Gatterer’s innovative propositions. Yet Adelung did
formulate a reservation, since he was not entirely convinced by the ahistorical
use of linguistic data propounded by the historian Gatterer:
Despite all of this, it is still not possible to draw from the mere kinship of languages,
without the contribution of history, any certain conclusion about a common
descent, however much Mister Gatterer (1771: 105) considers this sentence an
axiom not in need of any further proof. It is, after all, possible that one people is
forced to adopt the language of another, and history learns that this has happened
very often. How can we deduce in this case a single origin from a single language?⁸
In other words, Adelung objected that linguistic interrelationships could not be

the only criterion to stipulate a common origin for different peoples and empha-
sized the importance of history itself. Paradoxically, it was the grammarian
Adelung making this point in reaction to the historian Gatterer.
14.5 Conclusion
Even though historians and ethnographers generally took the conceptual pair for
granted when describing nations, regions, and past societies, a small number of
Mundarten von einander. Erstreckt sich die Abweichung auch auf andere als verwandte Hauptlaute,
und finden sich in den Ableitungen und Biegungen merkliche Unterschiede, so sind es verwandte
Sprachen, welche Verwandtschaft denn wieder ihre mannigfaltigen Stufen hat. Ganz verschiedene
Arten der Ableitung und Biegung, und größerer oder geringerer Unterschied in den Wurzeln und ihrer
Bedeutung geben mehr oder weniger verschiedene Sprachen’.
⁸ Adelung (1782: 243–4): ‘Bey dem allen läßt sich doch aus der bloßen Verwandtschaft der Sprachen
ohne Mitwirkung der Geschichte noch kein sicherer Schluß auf eine gemeinschaftliche Abstammung
machen, so sehr auch Herr Gatterer S. 105 des angeführten Werkes diesen Satz für ein Axiom hält,
welches keines weitern Beweises bedürfe. Es kann ja einem Volke die Sprache eines andern aufgedrun-
gen seyn, und die Geschichte lehret, daß solches sehr häufig geschehen. Wie läßt sich in diesem Falle
von einerley Sprache auf einerley Ursprung schließen?’
192    
them did engage in theoretical considerations of variable length on the question as

to how the language/dialect distinction could be put to use in their discipline. The
most fine-grained historiographical approach of the early modern era was devel-
oped by Johann Christoph Gatterer around 1770. The main principle on which
Gatterer relied was a well-known one: the close connection between a people and
its language and, more specifically, the way in which language could help one
reconstruct early human history. A fundamental assumption here was that
peoples speaking tongues that were claimed to vary only as dialects rather than
as entirely different languages had a common origin. Gatterer’s innovativeness
lies in the fact that he tried to make this idea more workable and concrete, while
grounding it in an objective mathematical approach in connection with the
concept of basic vocabulary. He thus developed a lexicostatistic method avant la
lettre, which became popular only much later, in the 1950s (see Chapter 20). His
ahistorical perspective on linguistic data, however, attracted criticism from
Adelung.
Gatterer’s stress on linguistic criteria in determining the status of a speech form
should no doubt be seen in connection with the rationalist trends of the eighteenth
century. In fact, his wide-ranging historiographical programme was inspired by
Enlightenment principles (Gierl 2012). Conceived as an ancillary science for the
study of history, Gatterer’s Sprachphilosophie, too, was informed by this intellec-
tual movement. His development of a statistical method for language study likely
constitutes the best proof of this. Framing language and dialect in an elaborate
method for language comparison, the Göttingen historian provided the most
rationalized early modern account of the conceptual pair.
15
Classes of variation
How do languages and dialects differ?
Albert Schultens and Johann Christoph Gatterer, although operating in different

contexts and disagreeing in several respects, exhibit three remarkable similarities
in their treatment of the conceptual pair. First, both scholars insisted on linguistic
properties in determining what distinguished a language from a dialect. This focus
on actual language betrays a desire to arrive at more objective, linguistically
motivated criteria. Second, they both attempted to systematize their discussions,
which also emerges from the schematized typographical setups of their texts.
Third, Schultens as well as Gatterer framed the language/dialect distinction within
a broader conceptual scheme—Schultens within a tripartite hierarchy and
Gatterer within a quadripartite division of the degree of kinship continuum. Are
these three similarities between Schultens’s and Gatterer’s ideas coincidental? Or
do they exemplify more general tendencies of the later stages of the early modern
era? I argue that they are indeed part of a broader trend, an idea which I develop
further in this chapter and the next. I first focus on the emergent interest in what
Schultens called the ‘classes of variation’, the ways in which related dialects were
believed to differ from one another, whether or not in opposition to distinct
languages.
15.1 Casting off John the Grammarian’s yoke
While it is true that eighteenth-century scholars such as Schultens and Gatterer

evinced a keen interest in the classes of variation, it does not mean that their
predecessors did not elaborate at all on them. What is more, in order to under-
stand eighteenth-century developments, I should briefly go back in time. The
respects in which the Ancient Greek dialects differed from one another had
already been systematized to a certain extent in the early Byzantine period. In
Chapter 2 (Section 2.3), I have indicated how John the Grammarian distinguished
three classes: the entire word, part of a word, and an accident of a word. Put
differently, the Greek dialects were claimed to differ in lexical terms, through letter
permutations, and through differences in accent and other diacritic marks. Since
John the Grammarian became one of the three canonical sources for information
on the Greek dialects, his framework was naturally adopted by Renaissance
DOI: 10.1093/oso/9780198845713.001.0001
194   
Hellenists, often literally and unquestioningly, which happened, for instance, in a

booklet on the Greek dialects dating to 1569 and published in France (Vuidius
1569: 138). Some philologists extended the framework so as to include other
classes of variation. The German professor of Greek Otto Walper (1589: 4–5), for
instance, added syntax to John’s three classes in his manual for the dialects aimed
at the students of Marburg university and first published in 1589. Theodore
Bibliander (1542: 52–3) likewise offered a general account of how dialects varied,
according to him on the level of an entire word, of part of a word, and of idiom.
His discussion was mainly inspired by John the Grammarian, but Bibliander
expressed his views when reflecting on the confusion of tongues at the Tower of
Babel rather than on the Greek dialects. As I have argued earlier (Chapter 5,
Section 5.6), however, his conception of the Latin term dialectus was highly
confused, and it is difficult to grasp how he related it to lingua.
During the sixteenth and seventeenth centuries, substantial systematic accounts
of the classes of variation typical of related dialects were principally restricted to
handbooks for the Ancient Greek language. From the end of the sixteenth century
onwards, however, scholars started to insert into their Greek manuals more
general accounts, which appear to have had a wider application and were no
longer dependent on John the Grammarian’s work. The German Hellenist Jakob
Gretser (1562–1625), for instance, who wrote the standard Greek grammar for
Jesuit colleges, defined dialectus as follows:
A dialect is nothing else than the characteristic and difference of one and the
same language, not employed by all using that language but only by some. Hence,
dialect is usually called “idiom” or “property”, whether this difference consists in
letters of the same word and affections of letters only, or in pronunciation only,
or in entire words, or in syntax and construction. The entire variation of dialects
indeed seems to be contained in these four parts.¹
Gretser distinguished in generic terms four different classes of variation among

related dialects: letters and their affections, pronunciation, entire words, and
syntax. John the Grammarian’s tripartite framework was, in other words, grad-
ually supplemented and eventually superseded in Renaissance Greek handbooks.
From about 1650 onwards, the classes of variation were extensively discussed
outside manuals on the Greek language and its dialects, too, which may be taken
as yet another symptom of the gradual de-Hellenization of the dialect concept.
This development made a start in works focusing on language origin in general
¹ Gretser (1593: 19): ‘D nihil est aliud, qua[m] vnius & eiusdem linguæ nota & differentia,
non omnibus qui illa vtuntur, sed aliquibus tantum vsitata; quorum ratione dialectus idioma seu
proprietas dici solet; siue differentia hæc consistat in solis literis eiusdem vocabuli, literarumq́[ue]
affectionibus, siue in sola pronuntiatione, siue in verbis integris, siue in syntaxi & constructione. His
enim quatuor partibus contineri videtur tota dialectorum varietas’.
.      195
and the so-called Oriental tongues in particular. As an example of the former may
serve the definition of Georg Stiernhielm, who opposed dialect to language in
terms of accidental and substantial variation in his preface on the origin of
languages. Stiernhielm linked his definition to linguistic features, too, as my
analysis in Chapter 12 suggests. In particular, Stiernhielm (1671: c.3v, d.4v) asso-
ciated language-level variation with differences in the lexicon and the roots of
words, whereas he believed related dialects to vary in terms of letters, morphology,
phraseology, lexical meaning, accent, and pronunciation. Stiernhielm’s acquaint-
ance Christian Ravis had expressed similar views in an earlier grammatical
publication on different Semitic tongues (see Chapter 12, Section 12.2).
According to Ravis (1650: 8, 44–5), related dialects varied mainly in terms of
accent, pronunciation, and accidents, whereas languages showed substantial dif-
ferences on the level of roots, the lexicon, and the meaning of words, which could
be designated with a modern phrase as lexical semantics. Like Stiernhielm, Ravis
linked this distinction to the Aristotelian interpretation of the conceptual pair,
which was no coincidence. As a matter of fact, the attention to classes of linguistic
variation intensified when scholars started to formulate the Aristotelian interpret-
ation in explicit terms (see Chapter 8, Section 8.1).
The focus on linguistic criteria may seem paradoxical in the light of the
increasing prominence of language-external interpretations of the conceptual
pair, most importantly in terms of status (see Chapter 9). It is, however, not
inconceivable that this linguistic turn in views on the conceptual pair was an
intuitive reaction against such subjective views on the language/dialect distinction.
Another plausible explanation is that scholars were developing an ever-increasing
interest in the specificities of language change and linguistic kinship. Philologists
decided, often on the basis of a superficial and intuitive analysis of select linguistic
forms and features, whether two speech varieties were related dialects or distinct
languages, a decision process with which language-external factors—social, polit-
ical, and ideological—frequently interfered. The precise mechanisms of such
language/dialect decisions deserve a closer analysis, which, however, lies outside
of the scope of this book.
15.2 From scattered comments to systematization
In the sixteenth and early seventeenth centuries, ideas about the ways in which
related dialects varied were primarily expressed in Greek handbooks. However,
they can also be gathered from scattered comments in other contexts. Let me treat
the three most important types of evidence. Firstly, some humanists discussed
how the dialects of a language other than Greek were different. The French
printer-scholar Henri Estienne, for instance, elaborated at length on French
variation in the outline of his book on the excellence of the French language.
196   
Estienne (1579: 138–46) assumed that French dialects, among which he included
for instance Savoyard, could differ from each other in terms of pronunciation,
lexical meaning, lexicon, and proverbial expressions.
Secondly, passing comments reveal a scholar’s assumptions on the classes of
variation associated either with language or with dialect. For example, the English
Jesuit Robert Persons (1546–1610) commented upon the events at the Tower of
Babel, following the authority of a number of pagan testimonies, that
if there had not bene some such miracle in the diuision of tongues; no doubt, but
that all tongues, being deriued of one, (as all me[n] are of one father,) the same
tongues vvould haue retained the self same rootes and principles, as in all
dialectes or deriuations of tongues we see that it commeth to passe.
(Persons 1585: 102)
This passage suggests that Persons believed languages to differ from one another
in terms of lexical roots and ‘principles’, which he assumed to be identical or
highly similar for related dialects. It is hard to catch the exact meaning of the term
principles here. Persons may have been thinking of primary forms, which would
make his phrase ‘rootes and principles’ tautological, or perhaps of grammatical or
syntactic rules.
Thirdly, some early seventeenth-century works contain utterly concise remarks
on languages and linguistic diversity, from which ideas on the classes of variation
can be inferred. Joseph Justus Scaliger, for instance, fleshed out his conceptual
distinction between language mother and offshoot dialect as follows: ‘the same
words effect that it appears as one language, but the transposition, change, and
modification of the same words create different offshoots’.² Scaliger thus suggested
that languages differed on the level of the lexicon, whereas related dialects
exhibited certain variations in individual words. He was no doubt thinking of
the age-old framework of letter permutations.
What exactly happened after around 1650 and especially in the eighteenth
century? Basically, language scholars systematized their views on the classes of
variation. The cases of Stiernhielm, Ravis, Schultens, Gatterer, and Adelung,
treated in the previous chapters, strongly suggest this. Numerous additional
examples could be cited from different parts of Europe and in different contexts,
usually in connection with the Greek, vernacular, or Oriental tongues. In such
discussions, the ways in which related dialects differed from one another usually
took centre stage, and the classes of variation among distinct languages were
treated only in the margin, if at all.
² Scaliger (1610a: 119): ‘eadem verba faciunt vnam linquam [sic pro linguam] videri: sed eorundem
verborum traiectio, immutatio, inflexio, aliam, atque aliam propaginem facit’.
.      197
Reading the many scattered remarks of early modern scholars together with the
more systematic eighteenth-century testimonies, one notices that there was no or
hardly any level on which related dialects were not believed to exhibit differences.
Instead of offering a detailed overview of the classes of variation associated with
related dialects and, less often, distinct languages, which runs the risk of being
confusing and uninformative, I limit myself here to some general observations.
A question that naturally poses itself is the following: why did early modern
scholars focus on classes of dialectal variation and not of language-level differ-
ences? Since they were unaware of this misbalance, one cannot but conjecture
about the reasons behind it. It might, for instance, be connected to the basic
assumption that linguistic variation was more or less unrestrained, which would
make it rather pointless to posit specific classes of variation on the level of
languages. Another pressing question concerns the vast number of classes of
dialectal variation scholars presupposed. How could two dialects be assumed to
differ in so many different ways and still be closely related? The answer must be
sought in the Aristotelian interpretation of the conceptual pair. Even though
related dialects differed from one another only on the surface, this did not mean
that they did not exhibit variation on many different levels. Despite the great
number of classes of variation suggested, which ranged from pronunciation,
accent, and letters through morphology, the lexicon, and semantics to syntax
and even idiom and style, dialectal differences were nonetheless prototypically
associated with only a restricted number of linguistic features, usually pronunci-
ation and so-called letter permutations. The superficiality of this type of variation
was often emphasized, but some scholars pointed out that there should be more
than a few letter differences in order to speak of a distinct dialect. The Dictionnaire
de Trévoux, for instance, an early eighteenth-century French dictionary which
enjoyed tremendous success, asserted that the presence of one type of vowel
contraction alone was not a reason to regard a speech form as a separate dialect
(see Dictionnaire de Trévoux 1721: .760).
The strong emphasis on letter variation was no doubt partly inherited from
Byzantine treatises on the Greek dialects, in which word modification processes
constituted the primary descriptive mechanism. This Greek framework is now
usually referred to as ‘pathology’ and was conflated with a similar Roman frame-
work known as the permutation of letters, as pointed out earlier (Chapter 2,
Section 2.3; Chapter 4, Section 4.4). It should be noted that the framework of
letter mutations was not exclusively used to describe differences on the level of
dialects. It was, in fact, of key importance for early modern approaches to
etymology as well as linguistic variation and change as a whole, both on the
level of language and dialect (see e.g. Cram 1999). Variation which modern
linguists would label ‘morphological’ was usually also accounted for by means of
such letter-change processes. As a result, early modern scholars tended to associ-
ate differences on the level of morphology with superficial linguistic variation,
198   
which accounts for the fact that before 1800 only a few intellectuals took differ-
ences in morphology to point to different languages. The English orientalist
Thomas Hyde (1636–1703), for instance, was exceptional in taking substantial
variation in the ‘grammatical part’ (pars grammaticalis) of a language, which for
him largely coincided with morphology, as a criterion to distinguish between
language and dialect. Hyde advanced this view in his Latin oration on the
antiquity of the Arabic tongue, made in 1691, when arguing in favour of a close
relationship between this language and Hebrew:
Out of said dispersion the Arabic language emerged, but its difference from
Hebrew was only dialectal, as is manifest also today both from the grammatical
part, in which the main criterion of languages is put, and from the very many
words which correspond to Hebrew words, as well as from the characters which
are derived from the Hebrew alphabet.³
The emphasis on letters rather than sounds can be explained by the close link
many scholars observed between a language and the writing system used to codify
it, an inheritance from antiquity. Before the modern period, the term littera was
understood in a double manner, designating both the letter itself and the sound it
represented. That is why scholars like Thomas Hyde involved writing systems in
determining the status of a variety. Scholars were, however, increasingly aware
that there was a mismatch between the spoken and the written word, an awareness
heightened by their engagement with dialectal variation. The Hellenist Erasmus
Schmidt, for instance, reminded his German audience in the early seventeenth
century that
we are now not talking about those dialects that are expressed by means of
different letters, but about those that are written with the same letters, but have a
different pronunciation. For instance, some Germans say ‘father’ as Vãter, with a
slightly longer a; others Văter with a slightly shorter a; still others Voter with an
o. And yet all write it with the same letter a, even though at the same time the
pronunciation is realized differently.⁴
Schmidt made this point in a discourse on the pronunciation of Ancient Greek,

held in favour of the traditional Byzantine pronunciation and against the new
³ See Hyde (1767 [1691]: 454–5): ‘Ex dictâ dispersione orta est Lingua Arabica, cujus quidem
differentia ab Hebraicâ erat tantum dialectica, uti & hodiè constat tam ex Grammaticali parte in quâ
præcipuum Linguarum Criterion ponitur, quàm ex plurimis vocibus quæ cum Hebraicis consonant; &
à Characteribus ab Hebraïco Alphabeto deflexis’.
⁴ Schmidt (1615: 239): ‘Non est nunc sermo de ijs dialectis, qvæ diversis literis exprimuntur; sed de
illis qvæ iisdem literis scriptæ, diversimodam pronunciationem habent. v. g. Patrem aliqvi Germani
dicunt Vãter, a longiuscula: aliqvi Văter, a breviuscula: aliqvi Voter per o: & tamen omnes scribunt
eandem literam a, qvum interim pronunciatio fiat diversa’.
.      199
ideas of Erasmus and others. It seems therefore that the debate on this topic raised
his awareness of the fact that letters could be pronounced differently and that
there could be an uneven relationship between speaking and writing, especially in
dialects.
Apart from pronunciation and letter mutations, the lexicon, too, was often
perceived as a major level on which related dialects differed from one another. The
lexicon as a level of dialectal variation was not only obvious from daily experience
but also had clear predecessors in Ancient and Byzantine Greek scholarship. Some
Greek philologists had already compiled collections of dialect words, either as a
tool for those wanting to write pure Attic or out of an antiquarian and philological
interest. As I have mentioned earlier in this chapter, John the Grammarian had
moreover included lexical variation among the three planes on which the Ancient
Greek literary dialects differed.
Since I have tried to offer a brief synthetic account of early modern ideas on the
classes of variation, I might have given the impression that these were largely
homogeneous. There was sometimes disagreement on this matter, however.
Christian Ravis, for instance, countered the widespread opinion that related
dialects differed from each other mainly in terms of pronunciation, since he
seems to have believed that variation on this level did not suffice to speak of
different dialects. That is at least what a passing statement of his suggests,
featuring in a discourse he held on the topic of the ‘Easterne Tongues’, in which
he discussed the easiness and wide spread of Hebrew: ‘I confesse there are many
different pronunciations, but never a Dialect, or Idiome in Africa’ (Ravis 1650: 71).
Ravis thus implicitly proposed a taxonomy with three layers: language – dialect –
pronunciation.
Direct debate among scholars about the classes of variation was, however,
almost entirely lacking. Yet scholars did react against general practices. The
Leipzig theologian and orientalist Bartholomaeus Mayer (1598–1631), for
instance, made the following objection to a group of unnamed scholars who
believed that the Hebrews, Aramaeans, and Syrians had one language:
On the contrary, I judge that the Chaldean and Syriac languages are not merely
different dialects of one and the same language, but that in the Babylonian
“speech confusion”, or certainly a little later, a peculiar language emerged,
I mean Aramaic, or Syro-Chaldean. This language, even though it approximates
more closely the nature of the mother than the other languages, is nevertheless
separated from it by such a distance that since it differs from Hebrew in terms of
letters, mode of inflection as well as very many essential primary forms, it should
be deservedly regarded as another and different language.⁵
⁵ Mayer (1629: 16–17): ‘E diverso statuo, linguam Chaldæam ac Syram non esse unius & ejusdem
linguæ diversas tantum dialectos, sed in γλωσσοσυγχύσει Babylonica, vel certè paulò post,
200   
From this observation it is clear that Mayer employed linguistic distance as his
main criterion to distinguish a dialect from a language in a discussion triggered by
predominant views on the interrelationships of certain Oriental tongues. Mayer’s
account is remarkable for two reasons. On the one hand, letter variation was
usually closely associated with dialect-level differences and not with distinct
languages, as in Mayer’s account. Morphological properties such as flexion were,
on the other hand, usually not distinguished from letter variation, but they were
by Mayer.
15.3 Two eighteenth-century outsiders in

the quest for linguistic criteria
Slight differences in terms of pronunciation, letters, and also the lexicon were
typically associated, often in very general terms, with dialect-level variation.
Towards the end of the early modern period, these categories frequently received
a more concrete interpretation, as the cases of Schultens and especially Gatterer
suggest. Two eighteenth-century scholars had the language/dialect distinction
even depend on one single linguistic feature.
In the first and earliest case, the criterion chosen is the most baffling, due to its
high specificity as well as its arbitrariness. In 1727, the German language scholar
Johann Georg Wachter (1663–1757) saw his German etymological dictionary
through the press in Leipzig. To his glossary Wachter added a preface on the
original Germanic language, in which he discussed among other things the origin
of language and linguistic diversity in general terms. Relying on the Bible, he
assumed like most early modern scholars that there had been one primeval
language, here unidentified, from which the others descended. He believed these
new languages to originally have been dialects of the original tongue which
gradually developed into distinct languages:
And so out of one single and primitive language (to which the sacred writings
and the agreement of nations testify) first various dialects were produced, and
almost as many variations as there were families of men or blood-relationships
among families. When these were subsequently separated, the dialects were
gradually transformed into languages. Later, out of these languages were formed
new languages, and dialects of languages, and languages of dialects, not suddenly
peculiarem extitisse linguam, Aramæam puto, sive SyroChaldæam, quæ licet propius accedat ad
matris indolem, quam reliquæ, ab ea tamen tanto intervallo distat, ut quia & literis, & inflexionis
ratione, & essentialibus insuper thematibus quamplurimis ab Ebræa discrepat, meritò pro lingua alia
ac diversa haberi debeat’.
.  -  . . . 201
and instantly, but gradually and step by step, with the help of time, space,
migrations, and colonies.⁶
In a footnote accompanying this passage, Wachter drily remarked: ‘I distinguish

languages from dialects in such a manner that difference of languages arises from
consonants, whereas difference of dialects emerges from vowels’.⁷
Even though Wachter formulated his highly specific criterion seemingly
independently, his idea was foreshadowed in the thought of the Flemish orientalist
Franciscus Raphelengius (1539–97). In a letter to the esteemed philologist Justus
Lipsius (1547–1606), dated 18 May 1584, Raphelengius commented on the close
relationship he saw between Persian and Dutch, observing: ‘I also find [in Persian]
dare for “door” (for deure [Dutch for “door”]), for I do not have a high regard of
vowels, since these can only form different dialects’.⁸ This highly specific criterion
might have been grounded in the idea that consonants made up the framework of
a word, into which vowels were implemented. Even though this consonant/vowel
criterion could strike certain modern readers as absurd, it is less so than it at first
sight appears. In fact, modern research has confirmed that consonant substitu-
tions often have a greater impact on the degree of intelligibility among speakers of
different language varieties than vowel changes (Gooskens in Boberg et al. 2018:
210). Raphelengius and Wachter did not, however, base their criterion on thor-
ough linguistic research but assumed its validity intuitively, without any concern
whatsoever about its workability.
The second case deserves more extensive attention, since it frames within a
debate over the status of a low variety vis-à-vis its roofing literary language. It
involves the Enlightenment thinker Ferdinando Galiani (1728–87), born in Chieti
in the then kingdom of Naples and today mainly known for his work as an
economist and humorist. Yet Galiani also authored a book on the Neapolitan
dialect (Del dialetto napoletano), first published in 1779. The booklet was highly
popular and incited ‘vivacious reactions’ (Brevini 1999: 1905–6), for instance by
his fellow citizen Luigi Sèrio (1744–99). Sèrio reacted against Galiani’s ennobling
of their native dialect and propagated a more plebeian variety of Neapolitan in a
work published a year later (Sèrio 1780). Copies of Galiani’s book sold out quickly,
and he made plans to reedit it with corrections and additions as early as 1780.
⁶ Wachter (1727: a.4–5): ‘Itaque ex una & primitiva Lingua (quod sacræ Literæ & Gentium
consensus testantur) suscitatæ sunt primo variæ , & totidem pene variationes, quot hominum
familiæ, aut familiarum cognationes(i), quibus deinde separatis, Dialecti paulatim abierunt in 
(ii). Ex his Linguis postea formatæ sunt novæ Linguæ & Linguarum Dialecti, & Dialectorum Linguæ,
non subito & repente, sed gradatim & pedetentim, temporum, locorum, migrationum, & coloniarum
adjumento’.
⁷ Wachter (1727: a.4–5): ‘(ii) Linguas a Dialectis sic distinguo, ut differentia Ling[uarum] sit a
consonantibus, Dialectorum a vocalibus’. On Wachter’s etymologicon see Jones (2001: 1106–7) and
Considine (2014: 135).
⁸ Raphelengius (1983: 123): ‘Invenio etiam “Dare” pro ostio (pro “deure”) nam non magni facio
vocales, cum harum sit tantum formare varios [sic] dialectos’.
202   
By his death, this work had been finished, but the augmented second edition
appeared only posthumously in 1789 (see Cortelazzo 1980: 112–13; Sansone 1999:
201–34; Pennisi 2009).
In his discussion of the Neapolitan dialect, Galiani proposed syntax as the main
touchstone for determining whether two varieties were ‘different languages’
(lingue diverse) or ‘dialects’ (dialetti). If the syntax was different, it concerned
distinct languages. If the syntax was the same or at least very similar, it concerned
different dialects of one language, despite any variation occurring on the level of
pronunciation, orthography, morphology, and the lexicon (see Galiani 1779:
esp. 33–4, cf. also 15, 27, 39). It is, however, not entirely clear what Galiani
meant by ‘syntax’, but his outline indicates that he was principally thinking of
phraseology and stylistics rather than word order or verb constructions. His focus
on syntax led him to regard Neapolitan as a dialect of common Italian, which was
based on the archaic literary Tuscan of Petrarch, Boccaccio, and Dante. In doing
so, Galiani wanted ‘to lessen the conflict between Tuscan, the prestige vernacular,
and local dialects such as those spoken by the educated classes of Naples and
southern Italy, by downplaying the phonetic and morphological differences and
emphasizing the common syntactic base’. In this context, he ‘proposed a moderate
standardization of southern dialects (starting with spelling) and a historically
founded description of Neapolitan which gave importance not so much to its
literary or folkloric uses as to its administrative uses’ (Pennisi 2009: 508).
In other words, Galiani saw Neapolitan as a dialect of the Italian language
because he did not want his native speech to supersede the Tuscan-based norm, at
least not in all genres of writing. Even though Galiani explicitly took syntax as his
main linguistic criterion to distinguish between a language and a dialect, he
nevertheless appears to have conceived of dialect as erratic, too, a view inspired
partly by his reading of Dante’s oeuvre and partly by his preconceptions about the
inferior position of Neapolitan vis-à-vis the Tuscan norm (Galiani 1779: 35, 55).
Galiani’s conception of the language/dialect pair was clearly not only grounded in
the linguistic feature of syntax but indirectly also in normative ideas about Italian
and other Italo-Romance varieties.
15.4 Conclusion
By 1650, scholars had teased out the most important interpretations of the
conceptual pair. The focus then shifted to rationalizing and systematizing existing
ideas. In fact, as I have tried to argue in this chapter, an objective turn occurred
after the mid-seventeenth century, which climaxed in the eighteenth century. This
objective turn manifested itself as a greater emphasis on linguistic criteria and on
the ways in which related dialects differed from one another as opposed to distinct
languages. In general, scholars seem to have believed that related dialects varied
.  203
primarily in terms of pronunciation, letters, and the lexicon, even though a wide
range of other linguistic levels recurred in discussions of the nature of dialectal
diversity. In this context, the superficiality of the variation was frequently
emphasized. Language-level variation, however, was closely associated with
radical differences in a twofold sense. On the one hand, scholars assumed that
distinct languages could vary on all possible levels, and these differences were
usually said to be of a substantial or fundamental nature. On the other hand,
language-level variation was claimed to manifest itself mainly in the roots of words.
The objective turn was partly motivated by rationalist Enlightenment trends, for
which the ideas of Johan Christoph Gatterer and Ferdinando Galiani can be cited
as prime examples. Paradoxically, this rational attitude did not necessarily mean
that the conceptual pair was recognized as potentially problematic; instead, most
scholars still perceived it as a self-evident common-sense distinction.
Early modern discussions of the classes of variation confirm the trend towards
de-Hellenization. Before about 1650, the nature of dialect-level variation was
mostly treated in manuals for the Ancient Greek dialects, which were for a large
part indebted to the three-class framework of the early Byzantine author John the
Grammarian. Afterwards, non-Hellenists took over the wheel, a change involving
a different attitude towards linguistic features in relation to the conceptual pair.
Numerous scholars started to look more, but far from exclusively, at language-
internal specifics rather than at language-external factors when distinguishing a
language from a dialect. Their proposals are sometimes surprising, as for instance
Galiani’s insistence on syntax, and occasionally even utterly particularistic, like
Wachter’s consonant/vowel criterion.
16
Between systematization and rationalization
The conceptual pair through
the Enlightenment lens
The period 1650–1800 was an age of reason, which took off in the wake of René
Descartes’s (1596–1650) work and culminated in the Enlightenment movement
during the eighteenth century. Rationalist trends went hand in hand with a
longing for the systematization of knowledge, evidenced by the popularity of the
genre of the encyclopaedia. It is indeed no coincidence that Diderot and
d’Alembert’s Encyclopédie is usually considered to be the ‘epitome of
Enlightenment literature’ (Mokyr 2002: 68). As I have tried to outline in the
previous chapters, this rational turn was reflected in ideas about the language/
dialect pair, too. Scholars such as Albert Schultens and Johann Christoph Gatterer
did their best to provide rational and systematic treatments of the matter in their
philological and historiographical programmes. They tried to offer objective
criteria, grounded in linguistic features. The symptoms of rationalization and
systematization go, however, much further. In what other ways do they manifest
themselves in seventeenth- and eighteenth-century understandings of the
conceptual pair?
16.1 Towards a dialectological tradition?
On 22 July 1702, the Saxon philologist Christian Gottlieb Schwartz (1675–1751)

and the even more obscure German scholar Abraham Helm presented to their
audience at Wittenberg university a Disputation on the causes of dialects, specif-
ically the Greek (see Van Rooy 2020c for a more detailed analysis). The text of
their dissertation, most likely authored primarily by Schwartz, the supervisor, was
printed in the same year. It offered the early eighteenth-century reader a well-
structured analysis of dialectal diversity as a general phenomenon. First of all, the
term dialectus and its numerous existing definitions were discussed. They were,
however, all rejected:
We leave the opinions of others behind us and add the following definition of our
own: a dialect is a peculiar change of a certain common language, especially in
DOI: 10.1093/oso/9780198845713.001.0001
.     205
terms of individual words, and that change is rather slight and different according
to the times and peoples using that same language.¹
Schwartz and Helm thus combined several interpretations of the conceptual pair.
Firstly, related dialects showed ‘slight’ differences as opposed to distinct languages.
Secondly, it was implied that dialects deviated from a common language. Thirdly,
dialect was associated with the entity of ‘people’, suggesting that language tran-
scended this level and encompassed different ‘peoples’. Fourthly, the reference to
time as a factor in linguistic diversification likely indicates that Schwartz and
Helm also took dialect in the language-historical sense of ‘descendant of a
language’, an interpretation which was to feature prominently in later chapters
of the dissertation.
In what followed, Schwartz and Helm (1702: .2–.2v) offered a kind of
synthesis based on earlier literature, elaborating extensively on the different
elements included in their idiosyncratic definition. The relationship of a dialect
to a common language, under which it was subsumed, was first treated. Then, the
authors advanced the idea that dialectal variation mainly manifested itself as
‘individual words that are altered in one way or another’, no doubt referring to
the traditional letter mutation framework, to which they added that variation on
the syntactic level occasionally also occurred.² In this context, Schwartz and Helm
(1702: .3–) took great pains to distinguish between the concepts of style and
dialect, frequently conflated by scholars because of the literary usage of the Greek
dialects (see also Chapter 5, Section 5.6). Next, they discussed the implications
of their claim that related dialects exhibited slight differences. Schwartz and
Helm continued by treating the different degrees of variation dialects could
display, which led them to posit a hierarchy of types of dialects; there were generic
dialects under which specific or local dialects resorted. Schwartz and Helm (1702:
.3–.1) then provided an overview of the causes behind linguistic change in
general and dialectal diversification in particular, stressing its gradualness. In the
few remaining pages of the dissertation, the authors moved from the general to the
specific and discussed the origin and peculiarities of the Greek dialects, as
announced by the title.
The disputation held by Schwartz and Helm may thus be considered the earliest
work entirely devoted to the problem of dialectal diversification per se, even
though they still had the reflex to focus strongly on the Greek context in the
final paragraphs. Yet they also alluded, albeit sparsely, to dialect-like variation in
German, Romance, and Hebrew. Their text was clearly an attempt at providing
¹ Schwartz and Helm (1702: .2): ‘Nos, missis aliorum sententiis, hanc nostram subiicimus
definitionem: Dialectus est Linguae alicuius communis, inprimis secundum singulas uoces, immutatio
peculiaris, eaque leuior, et, pro diuersis temporibus, ac populis illa eadem lingua utentibus, diuersa’.
² Schwartz and Helm (1702: .1): ‘uocibus singulis aliter atque aliter inflexis’.
206    
the reader with a comprehensive and systematic overview of the issue of dialectal
diversity. To this end, it treated key aspects related to the phenomenon, including
the criteria that should be used to determine the dialect status of a variety, the
classes of dialectal variation, the causes of linguistic diversification, and the
prototypical Ancient Greek context, knowledge of which was considered an
important philological skill in the early modern period, indispensable for anyone
desiring direct access to the great Greek authors of the past.
It is no coincidence that Schwartz and Helm’s disputation was presented at the
Wittenberg academy. This institution had a scholarly tradition of books and
dissertations on issues related to dialectal diversity. In the seventeenth century,
several works on the Greek dialects were published there (e.g. Schmidt 1604), as
were disputations on Oriental tongues (e.g. Pfeiffer and Martini 1663), often
conceived in terms of dialect. This tradition continued well into the eighteenth
century. Two other dissertations were presented in Wittenberg in the year 1709,
one on the Greek dialects (Thryllitsch 1709), and the other on the Greek Koine
(Kirchmaier and Thryllitsch 1709). These works, too, were systematizing treat-
ments, even though they were more narrowly focused on the Ancient Greek
context. Both disputations had the young Hellenist Georg Friedrich Thryllitsch
(1688–1715) as their presenter, and of at least one Thryllitsch was the sole author.
Towards the middle of the century, Wittenberg witnessed the publication of yet
another dissertation on the topic of dialectal variation, this time applied to the
context of the Oriental tongues and depending to a large extent on Albert
Schultens’s work (Groddeck and Treuge 1747; see Chapter 13, Section 13.4).
In 1782, this time in Berlin, the influential school reformer and late
Enlightenment scholar Friedrich Gedike (1754–1803) published an article entitled
On dialects, especially the Greek, which appeared in the first volume of the Berlin
magazine of sciences and arts. Much like Schwartz and Helm in 1702, Gedike
(1782) offered his readers a synthetic discussion of the phenomenon of dialects
with specific attention to Greek variation, while also referring to German, Latin,
and English diversity. However, very unlike Schwartz and Helm, he barely cited
any early modern sources and largely developed his ideas on his own. The topic of
dialects kept fascinating Gedike in later years, too. In 1794, he held a lecture on
German dialects, which he introduced in a manner similar to his 1782 contribu-
tion and in which he referred also to the Greek dialects (Gedike 1794). In fact, he
saw a close parallel between Greek and German dialects, comparing Ionic to Low
German, both smooth and delicate dialects in his perception. Doric was put next to
Upper German, both of which he claimed to be solemn and splendid. Between these
two dialects were the esteemed Attic and High German dialects, which approached
more closely the solemn Doric and Upper German dialects than Ionic and Low
German, respectively. Gedike repeatedly revisited this idea since his first formulation
of it in a 1779 work on linguistic purity (see Gedike 1779: –). This way, Gedike
could tie German dialects to their highly valued Greek counterparts. In fact, the main
.     207
goal Gedike tried to achieve with his 1794 lecture on the German dialects was to
promote the study of Low German (Plattdeutsch), not because it should replace High
German as the norm, but because it could be a source of richness and even correction
of High German. He even proposed to organize competitions for the collection of
dialect words (Gedike 1794: 332).
Most systematizing works were published in German-speaking areas. This
geographic concentration is not surprising, since Greek studies, which provided
the original incentive to study dialectal variation, never ceased to flourish there,
unlike in many other regions. These areas moreover witnessed an increasing
philological interest in the relationships among Semitic tongues, which most
scholars agreed were so closely related that they might well be dubbed ‘dialects’.
Last but not least, Protestantism caused a major concern over spreading the Word
of God in a unified and widely intelligible form of the German vernacular, which
seems to have made Protestant scholars more conscious of the vast variation in
their native speech. Yet systematizing treatises were not only produced in
Germany. There are examples of encompassing works from other regions as
well. For instance, the Swedish teacher and language scholar Sven Hof published
in 1772 a pioneering work on the dialect of Västergötland, an area in modern-day
Sweden (see Chapter 8, Section 8.5). His book consisted of a thorough grammat-
ical introduction and a long dictionary in which he compared Västergötland
dialect words to their Swedish equivalents. The intellectual circumstances in
which Hof produced his contribution resembled those of the German works
mentioned earlier.
Hof, too, was well-versed in the Greek language and its dialects; in fact,
scholarship on this tongue was a welcome model for Hof ’s description of the
Västergötland dialect (see Hof 1772: 12–13, 23). Additionally, Hof contributed to
establishing a norm for the Swedish language, as he wanted to arrive at a better
Swedish translation of the Bible and was concerned over Swedish orthography
(see Eriksson 1971; Droixhe 1978: 123–4; Hovdhaugen et al. 2000: 93–4). Like
Gedike after him, Hof considered dialects to be a source of good words that could
or even should replace foreign words. He moreover emphasized that knowledge of
the dialects was necessary in order to fully understand one’s native language, its
genius, and its orthography. This important idea indicates that Hof did not believe
dialects to be intrinsically corrupt. Indeed, they provided evidence for the nature,
essence, and history of the language to which they belonged. He nevertheless
suggested that a dialect should not be used in writing, claiming that the Swedish
variety he described actually should be unlearned rather than learned. What is
more, he wanted to teach his compatriots correct Swedish and claimed by way of
captatio benevolentiae that there was no glory in writing on vernacular dialects
(Hof 1772: 8–10, 28). Hof did, however, observe that dialects differed from each
other in terms of esteem: ‘all nations have some dialects they approve most, but
others they despise and condemn as if they were absurd’. Interestingly, such
208    
judgements were formed by the ‘imagination and usage of the ears’ rather than by
a rational decision.³
It should be clear that the phenomenon of dialectal diversity was gradually
becoming a topic in its own right after 1650, culminating in the eighteenth
century, when scholars such as Gedike and Hof insisted on the richness and
utility of regional dialects (cf. Droixhe 1978: 334–45). It did, however, remain a
kind of reflex to concentrate on the prototypical Greek dialects after a more
general discussion. Another symptom of the tendency towards systematization
and rationalization consists in the fact that it became somewhat more customary
to explicitly define language as opposed to dialect from the second half of the
seventeenth century onwards. This increased attention to the language pole did
not occur only in Georg Stiernhielm’s work and that of some of his readers but
also in several writings by followers of Albert Schultens (see Chapters 12–13).
It would nevertheless be premature to speak of a full-fledged dialectological
tradition in the early modern period, since the research field did not enjoy any
disciplinary autonomy, institutionalization, or other scientific organization, for
instance through specialized journals, as there was in the later nineteenth century.
The degree of dialogue between scholars was moreover limited. Be that as it may,
the term dialectology, more specifically in its Latin guise dialectologia, ‘study of
dialects’, was coined in this period as a composition of the Greek words diálektos
and lógos (λόγος), mirroring compositions like geologia, ‘study of the earth’. In
fact, dialectologia featured quite prominently on the title page of a work on the
Greek dialects and the way they allegedly figured in the Greek New Testament.
The work, entitled Sacred dialectology (Dialectologia sacra; Figure 16.1), was
published in Zürich in 1650 and compiled by the local professor of Greek
Caspar Wyss (1604/5–59).
The book was conceived as a manual to help readers understand the peculiar-
ities of the original language of the New Testament. However absurd and clumsy
this might seem to modern readers, the idea that features of all literary Greek
dialects could be found in the New Testament turned out to be a persistent one,
cultivated by many philologists in the early modern period and first systematized
by the Calvinist professor Georg Pasor in the first half of the seventeenth
century (see Van Rooy 2020a: Chapter 4 for a more detailed discussion). Pasor’s
Swiss colleague and follower Wyss coined the term dialectologia to strictly refer
to this subfield of Greek dialect studies. It was, however, soon applied to the
study of the Greek dialects in general, apparently first in a 1684 handbook for
Greek prosody by the poorly known Stuttgart teacher Johannes Bregius. The
³ Hof (1772: 14): ‘Etenim omnes gentes nonnullas habent dialectos, qvas maxime probant, alias
autem, qvas tamqvam insulsas fastidiunt & damnant; qva in re non tam rationis judicium, qvam
imaginationem & consvetudinem aurium seqvuntur’.
.     209
Figure 16.1 Caspar Wyss’s Dialectologia sacra, 1650

Source: e-rara. Public domain
210    
booklet was published in 1684 in Tübingen and had the title Greek prosody, with
dialectology, or Prosodia Græca, cum Dialectologiâ in the original Latin. This
designation caught on, and several German eighteenth-century Hellenists took it
over in this broader sense (see e.g. the title of Nibbe 1725). From here, it was only a
small step to apply it to vernacular dialects as well. Indeed, the Swedish scholar
Erik Benzelius the Younger (1675–1743) planned to produce a work entitled
Swedish dialectology (Dialectologia Suecica), containing word lists of Swedish
dialects, which never materialized, however (Considine 2017: 166, 173, 175).
In 1755, the term reappeared in the title of a brief description of the Hamburg
dialect, Hamburg dialectology or, in Latin, Dialectologia Hamburgensis; this trea-
tise was attached to the second edition of an extensive lexicon of the same dialect.
Both the description and the lexicon were produced by the Hamburg scholar and
poet Michael Richey (1678–1761), primarily out of a feeling of patriotism.⁴ An
additional motivation of Richey’s was dialect preservation, since he believed dialectal
features to be a source of linguistic richness (Haas 1994: ). For this reason, he
deplored the decline he noticed in the usage of his native dialect. What is more, at his
time of writing, even farmers were using High German words in an attempt at
passing themselves off as distinguished men (Richey 1755: –, ).
It comes as no surprise that Richey studied in Wittenberg around the turn of
the eighteenth century, at which time the topic of dialectal diversity was a hot
topic in Luther’s city. Even more unsurprisingly, Richey had himself specialized in
the Greek language and literature, too; he became professor of Greek and history
at the academy of his home city in 1717, holding both chairs until his demise (see
Waldberg 1889 for biographical information). Richey (1755: ) made explicit,
moreover, in the preface to the second edition of his Hamburg dictionary, that
lexica and descriptions of the Greek dialects had served as a source of inspiration
for his work. It is not unlikely that he adopted the term dialectologia from works in
this tradition, even though he did not cite any specific source in his lexicon.
Richey’s pioneering work was eagerly imitated across German-speaking
Europe, especially his lexicon but occasionally also his description of the
Hamburg dialect, his dialectologia. In fact, shortly after the appearance of the
second edition of Richey’s work, Johann Christoph Strodtmann (1717–56)
designed a brief Osnabrück dialectology (Dialectologia Osnabrugensis) and also
Germanized the term (Strodtmann 1756: , 3–8, 240). It seems that from about
this time the study of dialects had a name, dialectologia in Latin or Dialectologie in
its German form. The term remained in use to refer to descriptions of dialects and
their features, primarily in German-speaking areas, before becoming the designa-
⁴ Richey (1755). The first edition of the lexicon is Richey (1743). See Brekle et al. (1992–2005:
.172–5); Considine (2017: 117–19).
.     211
tion of a subfield of institutionalized linguistics in the late nineteenth century (cf.

Budzinski 2012). The surprising result is that this linguistic subfield received a
label before the rubric of linguistics was coined. This fact, too, suggests that a
dialectological tradition was emerging after 1650, which flourished in the eight-
eenth century, particularly in the Holy Roman Empire.
Despite the emancipation of the conceptual pair from the Greek heritage, the
systematization nevertheless started in the context of Greek philology. One could,
in fact, regard Schwartz and Helm’s 1702 disputation as having evolved out of
early modern manuals for the Ancient Greek dialects. The difference was, how-
ever, that the general discussion at the outset was not only given a wider applica-
tion, so as to concern the phenomenon of dialectal diversity as a whole, but also
constituted the dissertation’s main theme. This structure was then extrapolated to
philological studies of the so-called Oriental tongues. Systematization thus mainly
occurred within a philological context, with a primary focus on Greek and Semitic
diversity, even though book-length treatments of vernacular dialects began to
appear on a large scale, too. In these works, however, general conceptual issues
were usually not considered to be of direct concern, and authors mostly limited
themselves to ad hoc issues involving the description of a particular dialect.
16.2 The first sceptical voices
The rationalist attitudes adopted towards the conceptual pair in the eighteenth
century are not only noticeable in the linguistic turn occurring in definitions of
dialect versus language and in the emergence of an embryonic dialectological
tradition. Rationalist reflection also led a number of scholars to field questions
about certain usages of the language/dialect distinction and even about its ten-
ability tout court. Scholars were increasingly aware of the fact that the conceptual
pair actually covered different semantic oppositions, which was considered
undesirable. For this reason, some authors proposed to supplement the binary
language/dialect distinction with other concepts and to fit it into larger schemes.
I have argued in Chapter 13 that the Dutch orientalist Albert Schultens operated
with a tripartite constellation consisting of language, dialect, and degenerate
offshoot, thus adding a third element to the conceptual pair.
Schultens was certainly not unique in doing so. For example, some years before
him, the in his time highly popular French classical scholar and pedagogue
Charles Rollin (1661–1741), responsible for a revival of Greek studies in Paris,
had posited a tripartite scheme similar to Schultens’s:
These idioms or dialects of the Greek language are not like the different jargons
dominating in the different provinces of our France, which are an uncouth and
corrupted manner of speaking and which do not deserve to be called a language.
212    
Each dialect was a perfect language in its kind, which was in vogue with certain
peoples and which had its own rules and its own particular beauties.⁵
Rollin made a clear distinction between the cultivated varieties of Greek, termed
dialectes or idiomes, which enjoyed immense literary prestige, and the corrupted
patois of French, termed jargons, which he associated with the different provinces
of France.
The French grammarian Nicolas Beauzée (1717–89) designed a similar con-
ceptual scheme in his Encyclopédie article ‘Language’ (‘Langue’). In it, Beauzée
(1765: 249) linked language to nation, which embraced several peoples (peuples),
whereas he took dialect in the sense of ‘legitimate variety of a language, particular
to a people’, as opposed to patois, which he viewed as a degenerate variety of a
language spoken by the provincial populace of a unified state. Examples of dialects
could be found in German, Italian, and Greek. For patois, one should look at the
languages of the Roman Empire and the kingdom of France.
The scheme of Schultens, on the one hand, and those of Rollin and Beauzée, on
the other, are not entirely the same. In Schultens’s ideas, the historical dimension
and the process of language corruption played an essential role, whereas language-
external factors were of central significance for Rollin and Beauzée, who con-
nected their concepts of dialect and patois principally to cultural and political
circumstances. The very fact, however, that they felt the need to design such
tripartite divisions seems to indicate that they were no longer satisfied with a
simple binary conceptual opposition and that they wanted to arrive at a more
nuanced account of the omnipresent phenomenon of linguistic diversity.
Scepticism towards the conceptual pair and its polysemy manifested itself in
other ways, too. Some scholars insisted on differentiating between several inter-
pretations of the language/dialect distinction, since jumbling them together inev-
itably led to confusion. For instance, the Berlin-born orientalist Paul Ernst
Jablonski (1693–1757) published in 1714 a treatise on the mysterious Lycaonian
language, anciently spoken in Asia Minor and mentioned once in the New
Testament (Acts 14.11; on Lycaonian see e.g. Bruce 1992). The identity of this
language was much debated among early modern scholars. There were two main
camps, one arguing that it was a dialect of Greek, the other that it was a distinct
language. It was apparently first problematized in sixteenth-century England,
where the discussion was part of a larger disagreement between Catholic and
Protestant theologians, known as the Challenge controversy (Jenkins 2006:
115–54). In this debate, the practices of the Early Christian Church were the
⁵ Rollin (1726: 117): ‘Il n’en est pas de ces idiomes ou dialectes de la langue grecque, comme des
différens jargons qui regnent en différentes provinces de notre France, qui sont une maniere de parler
grossiére & corrompue, & qui ne méritent pas d’être appellés un langage. Chaque dialecte étoit un
langage parfait dans son genre, qui avoit cours chez certains peuples, qui avoit ses régles & ses beautés
particulieres’. On Rollin see the rich biography by Ferté (1902: esp. 16–17).
.     213
main topic of discussion, including the language used in Mass, where the issue of
the identity of Lycaonian came in.
Later, the matter became an autonomous philological topic discussed by Paul
Ernst Jablonski. In his Inquiry into the Lycaonian tongue, Jablonski discussed its
identity, as well as that of other obscure languages of Asia Minor such as Phrygian.
In this context, he reacted against the claim advanced by the Dutch historian
Theodorus Rijcke (1640–90) that Phrygian was a dialect of Greek, accusing his
predecessor of conceptual malpractice:
I am surprised that the most erudite man here takes two things, even though they
differ from each other, to be one. An offshoot of a language is one thing, a dialect
another. As a matter of fact, a language can depart from another language, of
which it is an offshoot, so far that it can no longer be held a dialect of it.⁶
Jablonski exemplified this last statement by referring to the case of the Latin
language. The general consensus was that it was an ‘offshoot’ (propago) of Greek,
but it could not be regarded as a dialect of it, since it had changed too considerably
from it (Jablonski 1714: 16–17). Jablonski thus clearly harboured suspicions with
regard to using the conceptual pair in a genealogical sense, proposing the distinc-
tion mother/offshoot as an alternative (cf. Chapter 10, Section 10.1). As I have
argued earlier (Chapter 10, Section 10.2), a similar concern had already been
voiced before Jablonski by another orientalist, the Swiss scholar Johann
Heinrich Hottinger.
Not all eighteenth-century scholars replaced the language/dialect distinction
with a multi-layer model. What is more, the fact that the conceptual pair had
different interpretations could still be a source of confusion and debate towards
the end of the eighteenth century. For instance, in this period, two English
scholars had an argument about the two concepts when discussing the language
of the Greek poet Homer. The classical scholar, traveller, and politician Robert
Wood (1717–71) stated the following in his Essay on the original genius and
writings of Homer, while refuting the views of ‘professed scholars and critics in the
Greek tongue’:
They compliment [Homer] for having enriched his language with the different
dialects of Greece; though the distinction of dialects can be only known to a
cultivated, and, in some degree, settled state of language, as deviations from an
⁶ Jablonski (1714: 16), with reference to Rijcke (1684: 465): ‘In quo illud miror Virum erudissimum
[sic] pro una re habere, quæ tamen inter se differunt. Aliud propago est linguæ, aliud Dialectus. Potest
namque lingua ab alia, cujus est propago, tam procul abire, ut pro ejus dialecto amplius haberi non
possit’.
214    
acknowledged standardb. — They point out his Poetical Licences; forgetting that,
in his time, there were no compositions in Prose.⁷
The historian Thomas Howes (1729–1814) responded to this observation of

Wood’s in the first volume of Critical observations on books, antient and modern:
Certainly Mr. Wood is mistaken in the first place, to assert that a distinction of
dialects can be only known to a cultivated state of language; for as dialects are the
derivatives and different modifications of one common original, so both these
derivations and the common original itself may be all equally barbarous and
uncultivated. ([Howes] 1776: 57)
To exemplify his genealogical conception of the distinction, Howes referred to the

Gothic language, from which he believed English, French, Dutch, German, and
Danish to have descended as dialects. He put this on a par with the Greek context;
out of the Pelasgian-Ionic common original language, several Greek dialects
developed. The synchronic perspective of Wood, in which standard language
was opposed to dialect, clashed with Howes’s language-historical interpretation
of related dialects as descending from one common ancestral language. Put
differently, Wood stressed the synchronic deviation of dialects from a language,
whereas Howes emphasized that the relationship between them is one of genea-
logical derivation. Their disagreement was, in other words, shaped by the different
ways in which they interpreted the conceptual pair.
By 1650 the main interpretations of the conceptual pair had been established.
The borders between them were, however, not always sharply drawn, as the
Wood/Howes case suggests. Some scholars tried to remedy this lack of precision
by clearly distinguishing between different understandings of the two concepts,
usually involving the language-historical interpretation, or by introducing a new
concept into their schemes. One author, the evangelical pastor Friedrich Carl
Fulda (1724–88), went even further and adopted an utterly sceptical attitude
towards the language/dialect pair, bringing its polysemy to a head. After studies
in theology, history, and statistics in Tübingen and Göttingen, Fulda developed a
keen interest in the German language. A late bloomer, he started his research on
this topic only around 1760. He wrote down his first results in an essay On the two
principal dialects of the German language, which won him the award of the Royal
society of sciences in Göttingen (Königliche Societät der Wissenschaften zu
⁷ Wood (1775: 238–9). Note b reads: ‘Nor would it be judicious to employ them indifferently. The
Bergamasc, Neapolitan, and Venetian dialects, do well on the Italian stage in the mouths of Harlequino,
Pulcinello, and Pantalone; but a Tuscan would never think of enriching his language by using them
promiscuously in an Epic Poem’.
.     215
Göttingen) on 9 November 1771.⁸ The essay provided a historical discussion of the

Germanic languages, their lexical roots, and the similarities in their grammatical
systems (Vogt 1974: 49). It was published two years later in Leipzig and included
at the back a table of Germanic roots.
The publication marked the beginning of an industrious period of philological
scholarship, during which he produced several lengthy works focusing on German
linguistic history as well as on its diversity. This activity lasted until his death in
1788, in which year Fulda brought out his Attempt at a general collection of
German dialect words. It is, however, his very first publication to which I should
return, as it was in this work already that Fulda expressed his doubts about the
value of the language/dialect distinction. In his essay, Fulda shrewdly observed:
Dialect and language are difficult to ascertain. Different ways in which one and
the same people speak a single language in different provinces in such a manner
that they nevertheless understand each other—that is very vague . . . If one
says that dialect is a slight distinction in a single word, language a complete
distinction of the word itself, then doubts are raised by the different, often
opposite meanings of the same word in a single or even different regions and
the daily rising and falling in value of the same word as well as, on the contrary,
the great resemblance among the words of apparently very diverse languages.
They are like blood relationship, which alienates itself in both lines, without
ceding to be common blood. Are only sisters dialects? Are mother and daughter,
aunt and niece no longer? One should call each language by its own name, and
kinship with a common name, and one should give up the word play.⁹
Although devoting only one paragraph to the matter, Fulda was one of the first,
if not the first, to cast doubt on the utility of the conceptual pair in such explicit
terms and to transparently point out its imprecise nature. He specifically men-
tioned the criterion of mutual intelligibility, which he dismissed as vague, and the
Aristotelian interpretation, which were often bracketed together by earlier
scholars, too (see Chapter 8). His suspicion was fuelled by the intrinsic variability
⁸ For biographical notes on Fulda see e.g. Franck (1878) and Püschel (2009). On Fulda’s essay see
e.g. Vogt (1974: esp. 49–52, 77–8, 84–6) and Brekle et al. (1992–2005: .175–8).
⁹ Fulda (1773: 36): ‘Dialect und Sprache sind schwer zu bestimmen. Verschiedene Arten eines und
eben desselben Volks, einerlei Sprache in verschiedenen Provinzen so zu reden, daß sie sich doch
gegenseitig verstehen, ist sehr unbestimmt . . . Sagt man, der Dialect sei ein geringer Unterschied von
einerlei Wort, die Sprache ein völliger Unterschied des Worts selbsten: so machen die verschiedenen,
oft entgegengesezten Bedeutungen desselbigen Worts in einerlei, auch verschiedenen Gegenden, und
der täglich steigende und fallende Werth desselben; wie im Gegenteil die grose Uebereinstimmung der
Wörter in offenbar höchstverschiedenen Sprachen, Zweifel. Sie sind wie die Blutsfreundschaft, welche
sich in beiden Linien entfernt, ohne daß sie aufhöre, ein gemeinsames Blut zu sein. Sind blos
Geschwister Dialecte? Sind Mutter und Tochter, Muhme und Nichte keine mehr? Man nenne jede
Sprache mit ihrem eigenen, und die Verwandtschaft mit einem gemeinschaftlichen Namen, und gebe
das Wortspiel auf ’.
216    
of human language in time, an intuition doubtless shaped by his language-

historical interests. The essence of language and linguistic variation, Fulda
believed, could not be captured in clear-cut and static concepts. In spite of his
reservations, he continued to use terms such as Dialect (later German: Dialekt), no
doubt for the sake of convention. In fact, a few pages further, Fulda (1773: 38) put
forward in a kind of Aristotelian reflex that, strictly speaking, variation should be
limited in order to speak of ‘dialects’, which differed from one another primarily
in terms of pronunciation.
Fulda did not cast doubt only on the conceptual pair but also on the widespread
metaphor of family relationship, which he took to be closely connected with the
language/dialect distinction. As an alternative, he proposed the use of specific
names to denote individual speech forms and a common label for the kinship
among different speech forms; he meant that it would be better to speak of
Germanic, which covered, among other languages, German, English, and Dutch,
rather than to associate these entities with abstract and metaphorical concepts such
as language and dialect, mother and daughter. For Fulda, in conclusion, the
distinction between these concepts was suspicious, since it was an all too simplistic
abstraction of the complex phenomenon of the variation present in human lan-
guage, which is always in flux. His scepticism did not, however, keep him from
extensively using the terms Sprache and Dialect throughout his linguistic oeuvre.
16.3 Conclusion
In the eighteenth century, general treatments of the phenomenon of dialectal

diversity appeared on the scholarly scene; these were concentrated mainly in
German-speaking Europe and seem to reveal a desire to systematize available
knowledge on the topic, very much in the spirit of the Enlightenment. This
systematization went hand in hand with the emergence of an embryonic dialect-
ological tradition, best evidenced by the publication of numerous monographs
devoted to the dialects of various languages, mainly Greek, vernacular, and
Semitic. In addition, the design of more nuanced conceptual schemes and scep-
ticism about the genealogical usage of the conceptual pair may be regarded as
symptoms of a wider eighteenth-century rationalist tendency. This reached a
culmination in Friedrich Carl Fulda’s work, where the inaccuracy of the concep-
tual distinction was explicitly problematized for the first time. However, the
question of the adequacy of the language/dialect pair came to be addressed at
much greater length only in the course of the nineteenth century.
V
FROM SILENT ADOPTION TO
OUTSPOKEN ABANDONMENT,
A F T E R 18 0 0
17
From Jones to Gabelentz
Silent adoption and renewed suspicion
I do not believe that the Greek, Latin, and other European languages
are to be considered as derived from the Sanskrit in the state in which
we find it in Indian books; I feel rather inclined to consider them
altogether as subsequent variations of one original tongue, which,
however, the Sanskrit has preserved more perfect than its kindred
dialects.
(Bopp 1820: 3)
This is how the Bavarian linguist Franz Bopp (1791–1867) imagined the relationships
between the ancient European languages and Sanskrit. They were dialects of a now
lost original tongue, best preserved in Sanskrit. Bopp put forward this statement in
1820 at the publication of his thoroughly revised English edition of his famous 1816
study of the Sanskrit conjugation system. This work rightly earned Bopp the title of
founding father of modern historical-comparative linguistics, and shortly after, in
1821, he obtained the chair of comparative grammar, the first of its kind, at the
recently founded Friedrich Wilhelm University in Berlin.¹ He was one of the first
scholars to thoroughly investigate Sanskrit from a linguistic point of view and
compare it meticulously with related tongues from Europe and Asia. His compara-
tive efforts laid the basis for an effective and sound methodology to prove the
kinship among the languages now known as ‘Indo-European’, a hypothesis first
formulated in the seventeenth century, when the protolanguage was usually
referred to as ‘Scythic’, and made popular by Sir William Jones’s (1746–94)
enthusiastic discourse to his Asiatic Society in 1786.
Jones, Bopp, and others such as Rasmus Rask (1787–1832) stood on the
threshold of a new era, in which linguistics finally emerged as a separate discipline,
emancipated from other fields of study, in the first place philology. One might
expect theoretical and methodological questions to be foregrounded in this period,
but was this indeed the case? And more specifically, did theorizing on the concep-
tual pair benefit from the development of institutionalized linguistics in the nine-
teenth century? Bopp’s remark quoted above suggests that the genealogical
¹ See Swiggers (2017: 177) for an assessment of Bopp’s principal merits and defects.
DOI: 10.1093/oso/9780198845713.001.0001
220    
interpretation of the language/dialect distinction rose to prominence in the context

of historical-comparative grammar. Can this observation be generalized?
17.1 The beginnings of modern linguistics
Even though Sir William Jones’s contribution to linguistics tends to be over-

emphasized by intellectual historians and linguists, it is impossible to overlook
him, if only because his work contributed to the emergence of the field as a
separate scholarly discipline. It did so by ‘put[ting] Sanskrit at the forefront of
European interest’ (Rocher 2001: 1160–1). Still, Jones was not yet a rigorous
linguistic scholar but rather a historian interested in fitting Asian ethnology into
the traditional framework of the Bible (Trautmann 1998: 106–7; Campbell 2006:
247). That he was not a linguist pur sang can be noticed in his varied usage of the
term dialect, among other things. He did not define the word, probably because he
considered its meaning common knowledge, like most scholars before him. Jones
employed it not only as a synonym of language and in the senses of ‘manner of
speaking’ and ‘spoken language’ but also to refer to dialect as stemming from a
language, much as it had been employed from about 1600 onwards.² The latter
usage comes as no surprise, as Jones’s interest in linguistic kinship was shaped
primarily by historiographical concerns. This focus was most famously embodied
in his hypothesis about a common ancestor for a wide range of tongues, which
included not only languages which linguists today consider Indo-European, such
as Sanskrit, but also tongues currently classified in other families, such as Egyptian
(Poser and Campbell 1992: 230; Blench 2006: 54). It is, however, remarkable that
the term dialect did not figure in the famous and oft-quoted passage of his Third
discourse in which he expressed this hypothesis (Jones 1799: 26).
In the early nineteenth century, too, the conceptual pair was taken for granted,
much as it had been in the early modern period. Franz Bopp, for instance, did not
see anything wrong with using the terms Mundart and Dialekt and the English
word dialect to denote a variety that was part of a group of genealogically related
varieties descending from a common ancestral language (see e.g. Bopp 1816: 10,
89, 97, 105, 116). Bopp’s Danish contemporary Rasmus Rask likewise used the
language/dialect pair throughout his published work without questioning it. What
is more, Rask cultivated certain usages of the concept of dialect very typical of
early modern scholarship. For example, he used a variation on the popular phrase
‘to differ only in dialect’ (Rask 2013 [1818]: 201). The German linguist and fairy-
tale collector Jacob Grimm (1785–1863) also used the pair in its language-
historical sense throughout his influential German grammar (Grimm 1819).
² For the former two meanings see e.g. Jones (1799: 80, 110, 138, 179, 201, 357). For the latter
meaning see Jones (1799: 60, 80, 121).
.    221
Put differently, Bopp, Rask, and Grimm relied on the conceptual distinction to
express, much like Jones, genealogical relationships among different tongues.
There were no theoretical discussions of the distinction in mainstream works by
these early linguists, no doubt because they initially focused on the actual lan-
guages, their interrelationships, and the Indo-European protolanguage rather than
on the metalinguistic concepts used to indicate the relationships among them
(cf. Morpurgo Davies 1998: 131–3 on Bopp).
The conceptual pair did, however, figure in Rasmus Rask’s ‘six degrees’ of
linguistic classification, a framework he developed only rudimentarily in two
texts of limited circulation. The first was a letter of 29 January 1819, the second
a manuscript treatise of unknown date. Rask presented his classification method
as his ‘new system for the grouping of all the languages on earth’, suggesting that
he created it ex nihilo. The ‘six degrees’ were, in the original Danish, Race, Klasse,
Stamme, Gren, Sprog, and Sprogart, ‘race—class—stock—branch—language—
dialect’. Rask considered this scheme necessary in order to avoid ‘get[ting]
completely lost in the infinite multitude of languages and dialects’.³ Rask did
not, however, formulate definitions of the different ‘degrees’ and focused in actual
practice on distinguishing between races.
17.2 Two pioneering dialectologists: Johann Andreas

Schmeller and Albert Giese
The early nineteenth century did, however, witness two scholars with an out-
spoken interest in dialectal diversity and the concept of dialect in particular. The
first, Johann Andreas Schmeller (1785–1852), is widely known among histori-
ographers and linguists.⁴ A pioneer of German dialect studies, Schmeller authored
the monograph The dialects of Bavaria (1821), which included a very early dialect
map, and composed a four-volume Bavarian dictionary (1827–37). The German
philologist treated Bavarian dialects principally out of an antiquarian interest and
assumed that dialectal diversification was caused in the first place by language-
external circumstances. Even though Schmeller (1821) did not explicitly define
dialect, readers of his work will notice that he believed dialect-level variation to
affect primarily pronunciation, morphology, and the lexicon.
More interesting still for early nineteenth-century conceptions of dialect is the
almost entirely forgotten Wittenberg-born linguist Albert Giese (1803–34), a
student of the German philologist August Boeckh (1785–1867) and Franz Bopp.
Giese discussed the conceptual pair extensively in his monograph on the Ancient
³ See Benediktsson (1980: 22–3), whose English translation I quote. See also Morpurgo Davies
(1998: 92).
⁴ See e.g. Brunner (1971); Scheuerer (1995); Niebaum and Macha (2006: 55–7).
222    
Greek dialects, which focused on Aeolic in particular and on the many myths
surrounding this obscure dialect. The work was published only posthumously in
1837, three years after Giese’s untimely death. His treatment of the language/
dialect distinction featured quite prominently in the book. Its first chapter, entitled
‘Language and dialect’, was almost entirely devoted to it, even if it started out as a
programmatic text about the aims and restrictions of historical-comparative
linguistics.⁵ Giese’s work has been described by a nineteenth-century biographer
as ‘one of the first and most inspiring attempts at applying the results and methods
of comparative linguistic research to Greek dialectology’.⁶
Not all nineteenth-century readers were, however, convinced of the quality of
Giese’s contribution. His fellow student August Friedrich Pott (1802–87), for
instance, argued that Giese’s work exhibited certain unscientific characteristics.
In particular, Pott (1974 [1884–90]: 92) took offence at Giese’s emphasis on Attic
as the main norm of the Ancient Greek language. Another, more striking criticism
was formulated by the presumed pioneer of modern Ancient Greek dialectology:
Heinrich Ludolf Ahrens (1809–81). Tellingly for the early nineteenth-century lack
of attention for metalinguistic concepts, Ahrens (1839: ) criticized his predeces-
sor Giese precisely for his lengthy introduction.
Giese used the term Mundart primarily in the genealogical, language-historical
sense. Greek, Latin, and Sanskrit were related dialects, each of which could help
the linguist complete the forms lacking in the others. Giese was aware that this
genealogical interpretation was extrapolated from the micro-level of a language
having different dialects to the macro-level of a language family and its members.
He moreover framed the two concepts within a ‘forever lasting change’
(immerwährende Bewegung), which, in his view, constituted one of the main
challenges for linguistic research (Giese 1837: 6).
Several pages later, Giese elaborated on a thorny issue: how could one distin-
guish between dialects of a language and related languages? He assumed that
dialects had a ‘closer connection’ (innigere Verbindung) than related languages,
but at the same time he observed that dialects of the same language could be highly
divergent, up to the point that a dialect resembled more a related language than a
cognate dialect. This realization led him to make critical observations on the
genealogical use of the term Dialekt and to elaborate on the ways in which the
concepts dialect and language should be distinguished in historical-comparative
linguistics:
However much the concepts ‘dialect’ and ‘language’ seem to flow into each other
in our prehistorically related language family, one can nevertheless offer certain
⁵ See Giese (1837: 3–22): ‘Sprache und Mundart’. The programmatic part coincides with pp. 3–7.
⁶ Leskien (1879: 151): ‘einer der ersten und anregendsten Versuche, die Resultate und die Methode
der vergleichenden Sprachforschung auf die griechische Dialectologie anzuwenden’.
.    223
points strictly distinguishing between these concepts. These are, on the one hand,
the original properties of a language and, on the other, certain more general
sound laws which arose within the autonomous development of a language, for
we find that both elements are preserved and observed in the dialects of a
language, but they can lack in a dialect of the related language.⁷
Giese summed up this idea by means of the following maxim: ‘that which connects
the dialects, keeps the prehistorically related languages apart’.⁸ He was thus one of
the first modern linguists, if not the first, to consciously adapt the conceptual pair
to the methodology of the new discipline of historical-comparative linguistics. He
was, however, aware that his analysis was tied to a specific time window, since
related languages such as Sanskrit, Greek, and Latin were originally themselves
dialects of one and the same language.
Giese (1837: 12) proceeded by applying his interpretation of the language/
dialect distinction to refute the age-old idea that Latin was a dialect of Greek or
a ‘mixed language’ (Mischsprache) composed of the Aeolic dialect and a number
of Italic languages. He pointed to, among other things, the Greek augment, not
present in Latin, the latter allegedly preserving the most ancient situation. In what
followed, Giese teased out the problem of the interrelationships of different
dialects and their relationship to the language of which they were part. He put
forward the following question in this context, which he did not manage to
answer: did dialects derive from an originally unitary language or were there
already dialectal differences in the beginning? Giese (1837: 13–14) did, however,
suggest that the latter scenario was possible, and that one should judge every
dialect relationship separately.
When discussing the place of related dialects vis-à-vis the language under
which they were subsumed, Giese characterized dialect as a relative concept
which had to be defined with respect to the overarching concept of language.
Interestingly enough, he used a traditional definition from Greek scholarship to
do so, clearly reading his own interpretation into it:
The following general bond is in the dialects: which they are not, as the language
is abstracted from the dialects as such, and the dialects are nothing else than the
́
language in its difference (idíōma glōssēs [“particularity of tongue”]). Yet the
dialects which have appeared in historical times are not the parts of the language
⁷ Giese (1837: 12): ‘So sehr auch bei unsrer urverwandten Sprachfamilie die Begriffe „Mundart“ und
„Sprache“ in einander zu fließen scheinen, so lassen sich dennoch bestimmte Punkte angeben, die jene
Begriffe streng auseinander halten; diese sind einerseits die ursprünglichen Eigenthümlichkeiten einer
Sprache, andrerseits bestimmte allgemeinere Lautgesetze, welche sich innerhalb der selbstständigen
Entwickelung einer Sprache bildeten; denn beide finden wir in den Mundarten einer Sprache bewahrt
und beobachtet, in einer der verwandten Sprache aber können sie fehlen’.
⁸ Giese (1837: 12): ‘was die Mundarten verbindet, hält die urverwandten Sprachen aus einander’.
224    
as of the whole, as far as we conceive this as the whole that existed ever since the
origin of the language, but they are the varieties of a specific language situation,
which, mainly determined by certain general sound laws, has emerged out of
another, older language situation . . . And the abovementioned general bond is
itself a variable one, as far as a language disperses more or less into dialects in
different periods.⁹
Giese tried to reconcile a synchronic interpretation of the conceptual pair with

its language-historical usage. This attempt was innovative, since early modern
scholars often made confused use of synchronic and diachronic conceptions of
language and dialect or, more rarely, argued to distinguish clearly between them.
There are other important differences between Giese’s views and early modern
thought, too. For instance, his insistence on a comparative approach to the dialects
of a language and on the difficulty of distinguishing between related dialects and
related languages seems to have no precedents. Even though the latter idea had
already been suggested by Fulda in the second half of the eighteenth century, early
modern scholars were usually not suspicious about the distinction between lan-
guage and dialect. As Giese probably had not read Fulda’s paragraph, it is safe to
conclude that he arrived at this conclusion on his own.
In what followed, Giese offered an alternative definition of dialect in which the
ethnic parameter was emphasized, once again clearly inspired by traditional Greek
definitions. Rather untraditionally, he linked it to the derivative nature of dialects
in relation to the language from which they stemmed (Giese 1837: 14, 17, 21). It is
remarkable that Giese, although operating with centuries-old interpretations of
the conceptual pair, did not refer to early modern works at all. This lack of
acknowledgement of earlier work was typical of nineteenth-century scholarship
in the humanities, which sought to break with the unscientific past, thus ignoring
early modern achievements.¹⁰
Outside Germany, there was also increasing attention to the study of dialects. In
France, in particular, the patois were mapped already in the first decades of the
nineteenth century by the efforts of members of the Coquebert de Montbret
family, especially Charles-Étienne and his son Eugène, both of whom worked at
the French Bureau de la statistique from 1806 onwards (McCain 2018: esp. 41,
⁹ Giese (1837: 14): ‘Jenes allgemeine Band ist in den Dialekten das, was sie nicht sind, denn es ist die
Sprache abstrahirt von den Dialekten als solchen, und die Mundarten sind nichts anderes als die
Sprache in ihrem Unterschiede (ἰδίωμα γλώσσης). Aber die in der geschichtlichen Zeit zum Vorschein
gekommenen Dialekte sind nicht die Glieder der Sprache als des Ganzen, insofern wir diese als das
Ganze, welches vom Ursprung der Sprache her bestand, auffassen, sondern sie sind die Arten eines
bestimmten Sprachzustandes, welcher, hauptsächlich durch gewisse allgemeine Lautgesetze bedingt,
aus einem andern, ältern Sprachzustande entsprungen ist . . . Und das genannte allgemeine Band ist
selbst ein veränderliches, je nachdem eine Sprache in den verschiedenen Zeiten mehr oder minder in
Dialekte auseinander geht’. Cf. Chapter 2, Section 2.2.
¹⁰ Cf. Lamers et al. (2020) and the other papers in the same journal issue.
.     225
78–82). As early as September 1807, Charles-Étienne developed the plan to

conduct enquiries into local speech varieties—mainly of Romance but also of
Germanic and other stock—which resulted in a thorough geographical study with
statistical components and the collection of texts in these regional varieties,
including numerous versions of the Parable of the Prodigal Son (cf. [Coquebert
de Montbret] 1831). They did not directly engage with the language/dialect
distinction, but apparently regarded the patois spoken in Napoleon’s empire as
vulgar, impure, and grammatically incorrect (McCain 2018: 78–9). In the spirit of
the recent French Revolution, French was propagated as the pure language of the
nation (see Chapter 23, Section 23.2).
In conclusion, remarkable though the early nineteenth century may be for the
history of linguistics in many respects, it is rather uninteresting as it comes to
theorizing about the language/dialect distinction. Active comparison of the Indo-
European languages and their specificities attracted most of the linguists’ atten-
tion, which left barely any space for more general reflections on methodological
and conceptual issues. The scholarly focus broadened in the last decades of the
nineteenth century, when dialectology emerged as a subdiscipline of linguistics
and the fixation with the ancient Indo-European tongues gradually ceded to a
more encompassing study of language. Several eminent linguists developed an
interest in the conceptual pair, which they approached with increasing scepticism.
Among them were August Schleicher and William Dwight Whitney.
17.3 Questioning the conceptual pair: August Schleicher

and William Dwight Whitney
At the outset of his 1861 comparative grammar of Indo-European languages, the

Thuringian linguist August Schleicher (1821–68) pointed out that ‘the distinction
between mundart, dialect, sprache cannot be determined in general terms’.¹¹ This
reservation did not, however, keep Schleicher from using these words throughout
his work. Rather unsurprisingly, he took them primarily in the long-established
language-historical sense. After all, Schleicher was a historical-comparative linguist
who tried to arrive at a better understanding of the Indo-European language family,
thus consolidating the young discipline of linguistics. The method he adopted for
doing so was as much unprecedented as it was daring; Schleicher relied on a
meticulous comparison of the Indo-European dialects to fully reconstruct their
common ancestral language, now usually known as ‘Proto-Indo-European’. He
even composed a short fable in his version of the reconstructed protolanguage.
¹¹ Schleicher (1861: 4): ‘der unterschid von mundart, dialect, sprache ist im algemeinen nicht fest zu
stellen’. On Schleicher’s linguistic views and his place in the history of linguistics see e.g. Robins (1967:
179–80), Koerner (1989); Morpurgo Davies (1998: 167–74, 196–201).
226    
The language-historical interpretation of the relationship between language

and dialect, moreover, fitted in well with Schleicher’s strict conception of lan-
guages as living organisms. In fact, it is not unimaginable that the family meta-
phors often used in connection with language families and language/dialect
relationships encouraged Schleicher in developing such a view on human lan-
guage, even though it was surely not as constitutive as his study of biology.
Schleicher did, however, assign other meanings to mundart and dialect as well,
associating it in particular with superficial linguistic variation. For instance, the
sound variation ‘:r = d’ in Umbrian, an Italic language, Schleicher (1861: 221)
characterized as ‘only dialectal’ (‘nur dialectisch’).
Two years later, in his essay Darwinian theory and linguistics, Schleicher (1863:
12–13, 19) consciously reflected on the language-historical interpretation of the
conceptual pair, transforming it into four layers of variation: languages, dialects,
subdialects, and idiolects. These he compared to concepts from his second-
favourite branch of learning: biology. Unlike biological diversity, however, lin-
guistic variation was of a gradual nature, Schleicher realized, which is why he
added that it was impossible for linguists to clearly distinguish between languages,
dialects, and subdialects. Furthermore, scholars greatly varied in their usage of the
terms Sprache and Mundart: ‘what some call languages, others call dialects and
vice versa’.¹² In conclusion, Schleicher was well aware of the arbitrariness of the
conceptual pair and the difficulties that presented themselves in defining and
applying it. This realization did not, however, keep him from resorting to the
conceptual pair in his language-historical work, doubtlessly for lack of any
satisfying alternatives.
A contemporary of Schleicher, the American linguist and Sanskrit specialist
William Dwight Whitney (1827–94) even devoted a lengthy article to the con-
ceptual pair. Whitney was the author of a contribution entitled ‘Languages and
Dialects’, which appeared in 1867 in The North American Review.¹³ The article
focused on ‘the phenomena of dialects, their rise and their overthrow,—the
diversification and assimilation of languages’ (Whitney 1867: 32). Working with
the concept of language community, Whitney proved himself aware not only of
regional variation in a language but also of the fact that individuals, families, and
professional groups exhibit differences of speech. In this context, he expressed a
view highly similar to Schleicher’s:
Now all these differences, limited as their range may be, are in their essential
nature dialectic; the distinction between such idioms, as we may properly call
¹² Schleicher (1863: 19): ‘So war denn begreiflicher Weise noch kein Sprachforscher im Stande eine
genügende Definition von Sprache im Gegensatze zu Dialekt u. s. f. zu geben. Was die Einen Sprachen
nennen, das nennen die Andern Dialekte und umgekehrt’.
¹³ See e.g. Silverstein (1971) and Alter (2005) on Whitney’s linguistic views.
.     227
them, and well-marked dialects, or related but independent languages, is one of

degree only, not of kind.¹⁴
It will be noticed that we have used the terms “dialect” and “language” inter-
changeably, in speaking of any given tongue; and it will also, we trust, have been
clearly seen how vain would be the attempt to establish a definite and essential
distinction between them, or to give precision to any of the other names which
indicate the different degrees of diversity among related tongues. No form of
speech, living or dead, of which we have any knowledge, was not or is not a
dialect, in the sense of being the idiom of a limited community, among other
communities of kindred but somewhat discordant idiom; none is not truly a
language, in the sense of being the means of mutual intercourse of a distinct
portion of mankind, adapted to their capacity and supplying their needs.
(Whitney 1867: 54)
Labels such as idiom, patois, dialect, and language were determined by ‘external
circumstances’ and, Whitney implied, not backed by linguistic evidence. Much
like Schleicher, Whitney drew a parallel with ‘natural history’ and its concepts of
variety and species. Despite claiming to use the terms language and dialect
interchangeably, it is clear that Whitney employed the term dialect to express
the genealogical kinship among English, German, and Swedish as well as among
Russian, Persian, and Hindi. What is more, he relied on the conceptual pair to
describe natural linguistic cycles; local varieties of a language developed into
dialects, which eventually diversified ‘into distinct, and, finally, widely dissimilar
languages’ (Whitney 1867: 38). This idea also implied that related dialects exhib-
ited only little variation, whereas distinct languages differed substantially from
each other.
Whitney regarded dialectal diversity as a universal phenomenon, ‘inseparable
from the being of any language at any stage of its history’ (Whitney 1867: 62). By
arguing so, he was reacting against two contemporary scholars, Ernest Renan
(1823–92) and Max Müller (1823–1900), who ‘affirm that the natural tendency of
language is from diversity to uniformity; that dialects are, in the regular order of
things, antecedent to language; that human speech began its history in a state
of infinite dialectic division, which has been ever since undergoing coalescence
and reduction’ (Whitney 1867: 57, see also pp. 60–1).
In other words, despite his claims that both terms could be used interchange-
ably, Whitney was clearly differentiating language from dialect. Indeed, his actual
usage suggests that he was limiting neither the term dialect to his social definition
nor the term language to his functional definition but also regarded them as two
poles of a conceptual bipartition with genealogical implications. Whitney’s
¹⁴ Whitney (1867: 36). Cf. Alter (2005: 140, 260–1) on Whitney’s relationship to Schleicher’s thought.
228    
variable reliance on the conceptual pair lays bare the polysemy of the language/
dialect pair and the flexibility with which linguists could use it. In Whitney’s case,
this variability is all the more remarkable since it occurred in one single text
entirely devoted to the issue of ‘Languages and Dialects’.
17.4 The late nineteenth century: between continued usage

and increasing scepticism
Around the time when the validity of the conceptual pair began to be questioned,
dialectology emerged as a separate subfield of linguistics, for which Georg
Wenker’s (1852–1911) dialect maps are usually cited as a foundational monument
today (see e.g. Chambers & Trudgill 1998: 15–16; Schrambke 2010). The results of
this new dialectological research seemed to confirm the arbitrary nature of the
distinction between language and dialect and led some linguists to argue that both
concepts were abstractions made from concrete linguistic facts and did not
correspond to any entities in actual reality (see Morpurgo Davies 1998: 238,
289; Chapter 18). Still, in the last decades of the nineteenth century, most linguists
kept on employing these abstract concepts in their work because it allowed them
to reach ‘empirically valid results’ (Morpurgo Davies 1998: 330).
It was especially the so-called neogrammarians, Junggrammatiker in German,
who made eager use of the dialect concept. This school of linguists which
originated at the University of Leipzig reacted against the narrow focus on the
ancient Indo-European tongues and emphasized the importance of dialect studies
in historical-comparative research, which, in their view, should be focused on
formulating infallible sound laws (ausnahmslose Lautgesetze) and turn the atten-
tion to actual living language (see Osthoff and Brugmann 1878: –; Murray
2010). Dialects were mined for their archaisms and, more importantly, served as
pure specimens of natural language development; the primary access gates to this
living lab of dialects usually were what linguists have later called ‘NORMs’, ‘Non-
Mobile Older Rural Males’ (e.g. Kristiansen in Boberg et al. 2018: 111). The
validity of this focus on NORMs later came to be criticized on various grounds,
including its gender bias and its assumption of lack of mobility and contact,
especially as transportation and communications technologies have since pro-
foundly changed human interactions.
In the second half of the nineteenth century, then, contemporary vernacular
dialects were, for the first time in history, considered an object worthy of serious
and systematic research in and of itself, the primacy of the written word finally
being superseded. Early dialect geographers were, however, not much concerned
over the distinction between language and dialect, even though their maps
foregrounded again the spatial conception of dialect. Neither did many neogram-
marians, who strongly disagreed with the dialect geographers on other points
.  229
(Jankowsky 1972: 133), discuss the problematic nature of the conceptual pair at
length. The neogrammarian Berthold Delbrück (1842–1922), for instance, still
took it for granted when drawing up his Introduction to language study (1880) and
interpreted it primarily in the usual language-historical sense (Delbrück 1880:
120–7). His colleague Hermann Paul (1846–1921), in turn, compared language
families, languages, and dialects to biological classes, much as Schleicher had
done, and emphasized the gradualness of linguistic variation (Paul 1880: 231–2).
This observation did not lead him to problematize the conceptual pair, however.
Towards the end of the nineteenth century, concern over the validity of the
concept pair increased. For instance, when discussing the precise number of
the world’s languages, the linguist August Friedrich Pott (1974 [1884–90]: 251),
a student of Franz Bopp’s, pointed out that it was difficult to observe a clear
distinction between language and dialect. A more interesting example still is the
German sinologist Georg von der Gabelentz (1840–93), who reflected on the
criteria used to distinguish between language and dialect in his work Language
science, its tasks, methods, and current results, which was first published in 1891
and appeared in an enlarged edition posthumously in 1901. Even though
Gabelentz’s work was relatively uninfluential, it is worthwhile to take a closer
look at his interesting comments on the conceptual pair, since they illustrate very
well how it was increasingly questioned in the late nineteenth century.¹⁵
According to Gabelentz, the only correct criterion whereby to distinguish a
language from a dialect was mutual intelligibility, which, however, had not yet
been defined in an objective fashion. It was this flaw which he tried to remedy.
Gabelentz endeavoured to do so by insisting on the importance of the speech of
uneducated people. Although attempting to design a workable criterion, he was
nonetheless aware that it was impossible to satisfactorily solve its arbitrariness. For
there was no freezing or boiling point in mutual intelligibility, as there was with
temperature. As a consequence, the language/dialect distinction and the criterion
he proposed could only be used for convenience. Still, such distinctions did have
scientific value, claimed Gabelentz, even though he refrained from explaining in
what this value exactly consisted.
17.5 Conclusion
The conceptual pair was silently adopted by linguists when the study of language
emerged as an autonomous discipline in the early nineteenth century, as if it were
an ever-present and unquestionable given. The language/dialect distinction was
not recognized for what it was: a product of the reflection and linguistic experience
¹⁵ Gabelentz (1901: 54–8). On the limited impact of Gabelentz see Elffers (2008), but compare this
with Albrecht (2007: 22–3, 178).
230    
of Renaissance scholars. The language-historical interpretation was foregrounded

at an early stage, obviously connected to the focus of institutionalized language
studies in the nineteenth century, largely coinciding with historical-comparative
Indo-European linguistics. The early comparative philologist Albert Giese was
exceptional among his colleagues in that he consciously attempted to restyle the
conceptual pair to the principles of the young discipline. Tellingly, Giese did not
elaborate on early modern views but tried instead to connect his ideas to trad-
itional Greek thought. Interpretations other than the language-historical sense
lived on, too, principally mutual intelligibility and the Aristotelian criterion. The
spatial conception of the distinction was likewise rehabilitated when dialect
geography emerged in the late 1870s and became the primary dialectological
paradigm for decades to come.
In the second half of the nineteenth century, the conceptual pair was increas-
ingly problematized by prominent linguists such as Schleicher, Whitney, and
Gabelentz. It comes as no surprise that they were not aware of Fulda’s scepticism,
as there seems to have been a general neglect of early modern work. It is, however,
remarkable that they refrained from referring to one another’s ideas when dis-
cussing the conceptual pair. This lack of debate stands in stark contrast with their
treatments of other issues, in which they displayed detailed familiarity with the
work of their colleagues. It appears that in their historical-comparative enthusi-
asm they had no use for theoretical debates on key concepts such as language and
dialect, which they initially perceived as obvious auxiliary concepts. Or perhaps
they wanted to keep themselves far from this thorny problem, which could shake
the very foundations of linguistic study. After all, if it was unclear what constituted
a language or a dialect, the very object of this new discipline would be in danger.
Yet linguists could not look away for ever.
18
Schuchardt the iconoclast
In her magisterial overview of nineteenth-century linguistics, the late Anna

Morpurgo Davies devoted a section to ‘Hugo Schuchardt, language and dialect’.¹
She did so with good reason, as the German linguist Hugo Schuchardt
(1842–1927) radically reacted against the common practice among his fellow
linguists to take the concepts of language and dialect for granted. What led
Schuchardt to formulate a smashing critique of his colleagues and their funda-
mental assumptions? To answer this question, I have to retrace my steps to the
nineteenth-century debate on the modality of linguistic diversification, centring
around the question: is it a branching out or a diffusion in waves (cf. Goebl 1983)?
18.1 Tree or waves?
Even though August Schleicher on several occasions expressed his discontent with
the arbitrary language/dialect distinction, he had done so only in passing and in
the end kept on working with it, most likely for practical considerations. In fact,
Schleicher needed it to develop his influential Stammbaum theory. According to
this organicist model, the original Proto-Indo-European language ramified into
distinct families of languages. These families in turn split up into different
languages, which finally branched out into different dialects (Figure 18.1).
Schleicher’s tree model, in other words, presupposed strict boundaries between
linguistic branches, languages, and dialects. It is well known that his student
Johannes Schmidt (1843–1901) reacted against the rigidity of Schleicher’s divi-
sions with his wave theory (Wellentheorie). In his model, Schmidt tried to stress
the connection between the different Indo-European languages rather than their
division. They were initially a continuum of dialects in which some gained the
upper hand, thus abolishing adjoining varieties and creating larger gaps in the
continuum. The Attic dialect, Schmidt (1872: 27–8) pointed out, eclipsed all other
Greek dialects, becoming the main variety of Greek. To clarify his hypothesis,
Schmidt employed the image of a slanted line to describe the initial situation,
which became indented when gaps were created in the linguistic continuum as
certain tongues came to prevail—contrast Figures 18.2 and 18.3.
¹ Morpurgo Davies (1998: 287–90), an important source of inspiration for the present chapter.
DOI: 10.1093/oso/9780198845713.001.0001
232   
Figure 18.1 August Schleicher’s (1863) family tree model

Source: Archive.org. Public domain
Wave theory, however, has taken its name from another powerful image of
Schmidt’s, developed only very briefly a few sentences earlier as an alternative to
Schleicher’s Stammbaumtheorie. Linguistic differentiation was like a wave,
‘extending itself by means of concentric circles which become ever weaker as
they are further removed from the middle’.² Schmidt tried to develop a model that
better matched the complex reality of language diversity and change than the
artificially strict divisions inherent to his teacher’s tree model. The alternative
models divided linguists into two camps, even though some, like August Leskien
(1840–1916), insightfully argued at an early stage that the theories complemented
rather than contradicted each other (Leskien 1876).
The debate shook linguistics to its foundations—finally, one might even say, as
for a long time theoretical issues had been blatantly neglected. One of the major
problems that now compelled the attention of linguists was the language/dialect
distinction. The basic assumption that ‘languages or dialects were self-contained
² Schmidt (1872: 27): ‘Ich möchte an seine stelle das bild der welle setzen, welche sich in concen-
trischen mit der entfernung vom mittelpunkte immer schwächer werdenden ringen ausbreitet’.
.    233
Sanskrit
Celtic
Figure 18.2 The linguistic continuum as a slanted line according to Schmidt

Source: Schmidt 1872: 27–8
Sanskrit
Celtic
Figure 18.3 The linguistic continuum broken up according to Schmidt

Source: Schmidt 1872: 27–8
units easily identifiable’ was increasingly contested, not in the least by Hugo
Schuchardt, who even before Schmidt published his book in 1872 had pointed
out the impossibility of adequately classifying languages and dialects and went
much further than his colleague, who had still believed in an originally unitary
Proto-Indo-European tongue (Morpurgo Davies 1998: 285–7).
18.2 A pioneering outsider
Who was this Hugo Schuchardt (Figure 18.4), who adopted an unprecedentedly
radical attitude in studying human language?³ Schuchardt was born on 4 February
1842 in the Thuringian town of Gotha, where he spent most of his youth. After
passing his final exams at the local gymnasium in 1859, he moved to Jena to study
law, but he soon changed to philology, attending the classes of, among others,
August Schleicher, the man whose Stammbaumtheorie he later criticized. In 1861
he switched to the University of Bonn, where he obtained a PhD in 1864 with a
dissertation on the vowel system of Vulgar Latin, written in scientific Latin. In
1866–8 he published a three-volume German version of this work, which
³ Biographical information is drawn from the detailed and useful website of the Hugo Schuchardt
Archiv, consultable at http://schuchardt.uni-graz.at/ (last accessed 14 January 2020). For Schuchardt as
outsider see e.g. Swiggers (2000).
234   
Figure 18.4 Hugo Schuchardt in 1922, picture by Hermann Urtel

Source: Hugo Schuchardt Archiv, Graz. Reproduced with permission
remained his most extensive publication. In 1870, Schuchardt obtained his habili-
tation at the University of Leipzig with a study of Romansh sound changes,
accompanied by a notable trial lecture, ‘On the classification of the Romance
dialects’. By this point, he had already developed several of his main fields of
interest: Romance philology and contact linguistics, both areas in which he broke
new ground.
In 1873, Schuchardt was appointed professor of Romance philology at the
University of Halle, a neogrammarian stronghold. Three years later he managed
to obtain the chair in the same discipline at Graz University on the recommen-
dation of Johannes Schmidt. From the second half of the 1870s onwards,
Schuchardt undertook several research trips with various destinations across
Europe. His travels to Wales and Spain drew his attention to new languages,
and he started working on Celtic in the 1870s and Basque in the late 1880s. Also in
the 1880s, he quasi-founded a new linguistic subfield, that of creole linguistics,
which for Schuchardt involved studying the mechanisms behind what he typically
called ‘language mixture’, or Sprachmischung in German.
.     . . . 235
In 1885, he wrote a radical critique of the neogrammarians, reproaching their

too systemic and mechanistic approach to language change. By this time,
Schuchardt had become a well-known linguist, who corresponded with countless
researchers across Europe.⁴ His fame and his extensive scholarly network he owed
to his many travels and his pioneering contributions to linguistics. In the 1890s,
Schuchardt developed an interest in still other languages, including Caucasian
tongues and Iberian, an extinct language anciently spoken on the Iberian penin-
sula, which he believed to be closely related to Basque. Schuchardt tried to prove
this controversial hypothesis rather unsuccessfully.
In 1900, Schuchardt retired, but this did not mark the end of his scholarly
career. He kept on working on language-specific themes, especially regarding
Basque, while broadening his perspective by touching on more general linguistic
issues, too. In 1922, his Austrian admirer Leo Spitzer (1887–1960) edited the Hugo
Schuchardt-Brevier: Ein Vademecum der allgemeinen Sprachwissenschaft, an
anthology of Schuchardt’s work compiled in celebration of his 80th birthday.
Copies of this book sold out quickly, and Schuchardt collaborated with Spitzer to
produce a new edition, corrected and enlarged by the author himself. He did not
witness its eventual publication in 1928, however, as he had passed away on 21
April 1927.
18.3 Breaking down the walls between languages and dialects
A student of Schleicher’s, Schuchardt was thoroughly acquainted with his profes-

sor’s work. He could even appreciate it, and in the first volume of his book on the
Vulgar Latin vowel system, published in 1866, he applauded Schleicher’s (1863:
19) observation that ‘what some call languages, others call dialects and vice versa’
(see Chapter 17, Section 17.3). Schuchardt believed, however, that this needed a
further qualification, all too long neglected by fellow linguists:
This lies in the essence of language diversity, which is a gradually growing one.
Untermundart, Mundart, Dialekt, Sprache are relative concepts. If we wanted to
argue that related dialects were promoted to the rank of languages on condition
that mutual understanding ceased, then nothing would be gained by that; for the
ability of understanding is different in different persons and understanding itself
is endlessly gradual.⁵
⁴ See the online Hugo Schuchardt Archiv for an overview and edition of his vast correspondence.
⁵ Schuchardt (1928: 164): ‘Es liegt dies im Wesen der Sprachdifferenz, die eine allmählich wach-
sende ist. Untermundart, Mundart, Dialekt, Sprache sind relative Begriffe. Wollten wir behaupten,
verwandte Dialekte stiegen dann zur Würde von Sprachen empor, wenn das gegenseitige Verständnis
aufhörte, so würde dadurch nichts gewonnen sein; denn die Fähigkeit des Verstehens ist bei verschie-
denen verschieden und das Verstehen selbst unendlich abgestuft’,
236   
Schuchardt was, in other words, the first to indisputably recognize the complex
and gradual nature of linguistic variation, which, he believed, arbitrary concepts
such as language, dialect, and subdialect could never adequately convey. For this
reason he also rejected mutual intelligibility as a workable criterion.
Whereas his illustrious professor had not fully realized the implications of his
observations on these key concepts, Schuchardt did not hesitate to problematize
them even at the outset of his career—he was only twenty-four when his first
volume on the Vulgar Latin vowel system appeared. In the third part of the same
work, published in 1868, he further reflected on the gradual nature of linguistic
variation, observing that ‘neighbouring Dialekte, Mundarten, Untermundarten
etc. are not abruptly delimited from each other but approximate each other,
flow into each other’.⁶
Schuchardt considered it impossible to draw strict borders between speech
forms, which he believed to be a consequence of contact phenomena. Different
language regions were connected because speakers of different varieties interacted
with members of other speech communities. The intensity of this contact differed
depending on the degree of civilization. This contact was, in other words, a
gradational phenomenon itself. Schuchardt tried to visualize his conception of
dialectal diversity by means of a telling, although rather complicated scheme
(Figure 18.5), which he explained as follows:
If in the accompanying figure the four large triangles represent main dialects and
the sixteen small ones represent subdialects, then e.g. a, A, α, a are not to the
same degree different from the dialects having B as their centre, but a to the
greatest degree, A to a lesser degree, α, a to the smallest degree. What is more, the
general similarity between α and b is possibly greater than that between α and a.⁷
a α
A
δ b
a
D d b B
c
d β
C Figure 18.5 Schuchardt’s (1928 [1868]: 165) conception of
γ c
dialect groups visualized
⁶ Schuchardt (1928: 164): ‘so finden wir fast überall, dass benachbarte Dialekte, Mundarten,
Untermundarten usw. nicht schroff gegeneinander abgrenzen, sondern sich aneinander annähern,
ineinander überfliessen’.
⁷ Schuchardt (1928: 164–5): ‘Wenn in der nebenstehenden Figur die vier grossen Dreiecke
Hauptdialekte, die sechzehn kleinen Unterdialekte darstellen, so sind z. B. a, A, α, a nicht im gleichen
Grade von den Diall. mit dem Zentrum B verschieden, sondern a am meisten, A weniger, α, a am
meisten [sic pro wenigsten]. Ja, die allgemeine Ähnlichkeit zwischen α und b ist möglicherweise grösser,
als die zwischen α und a’.
.     . . . 237
Yet the schedule was still too rigid in its division of speech forms, Schuchardt
believed, which was why he resorted to another image, one which sounds strik-
ingly familiar:
Let us imagine language in its entirety as a body of water with an even surface;
this water is put in motion as a result of wave centres that are formed at different
locations in it, the systems of which, depending on whether the intensity of the
driving force is of a greater or smaller extent, traverse each other.⁸
Schuchardt wrote this in 1868, four years before Schmidt developed this image in
his booklet on the kinship relationships among Indo-European tongues.
Apparently, Schmidt did so independently from his colleague, since on 14 April
1874 he asked Schuchardt in a letter to provide him with the reference to the work
in which Schuchardt had arrived at insights similar to his own.⁹ It seems that
Schuchardt himself had drawn Schmidt’s attention to their parallel conclusions
after a presentation by Schmidt at a philology conference in Leipzig in late May
1872, where they probably met for the first time. Even though Schuchardt referred
to his 1870 trial lecture on Romance dialects rather than to his magnum opus on
Vulgar Latin, the fact that Schmidt had to make an enquiry about Schuchardt’s
writings suggests that he was not well-acquainted with his colleague’s work.
It is not very surprising that Schuchardt noticed a parallel between Schmidt’s 1872
treatise and his own lecture on the Romance dialects. In these works, both linguists
emphasized the difficulty and even absurdity of drawing exact borders between
languages and dialects; both linguists criticized Schleicher’s Stammbaumtheorie,
pointing out the unevenness of linguistic change and the asymmetric cohesion
between speech forms. Schmidt’s request to have Schuchardt’s lecture in a printed
form was in vain, however, since Schuchardt only published it in 1900. He eventually
decided to do so because he felt embarrassed, as he had cited this unpublished lecture
so very often that he could no longer leave it unprinted. He was moreover aware of
the fact that, in spite of its blemishes, its content was pioneering. In this lecture,
Schuchardt had argued against the classificability of dialects in general; he spoke in
German of ‘die Nichtklassifizirbarkeit der Mundarten’ (Schuchardt 1900: 1*). In
other words, he had pointed out the inadequacy of concepts such as dialect because
of contact and mixture phenomena to which all speech forms were susceptible, albeit
⁸ Schuchardt (1928: 165): ‘Denken wir uns die Sprache in ihrer Einheit als ein Gewässer mit glattem
Spiegel; in Bewegung gesetzt wird dasselbe dadurch, dass an verschiedenen Stellen desselben sich
Wellencentra bilden, deren Systeme, je nach der Intensivität der treibenden Kraft von grösserem oder
geringerem Umfange, sich durchkreuzen’. Cf. Goebl (1983: 33); Mayrhofer (1983: 135–6); Pirazzini
(2013: 29–30).
⁹ See letter 01-10093 at the Hugo Schuchardt Archiv.
238   
to different degrees.¹⁰ By publishing this old text, Schuchardt wanted to demonstrate

his own originality on this point. The fact that he still found it relevant to divulge his
lecture so many decades after he had written and read it may suggest that in 1900 the
discussion about key metalinguistic concepts such as language and dialect was
intensifying.
18.4 Linguists, shibbolethists
Schuchardt adopted a more radical attitude than Schmidt, as he argued that names
for languages and dialects were abstract collective terms taken from geography
rather than identifiable linguistic entities.¹¹ In fact, there were even differences
within cities such as Venice and Rome, he claimed, thus consciously reiterating a
centuries-old topos (Schuchardt 1900: 12–14; cf. Chapter 7, Section 7.2). In later
publications, Schuchardt dwelt in greater detail on the problematic nature of
concepts such as language and dialect. Especially in a treatise on learning the
Basque language in its Sara variety, published in 1922 and not included in the
Brevier, he developed interesting ideas on the term Mundart, even if it seems that
he did not want to lose himself too much in theoretical considerations.
When discussing the different accentual systems within Basque and their
dialectal mapping, Schuchardt felt compelled to broach ‘the general problem of
dialects’ and to touch on the troublesome word Mundart, despite his strong
aversion to terminology—nomina sunt odiosa, he complained.¹² Mundart was
not a scientific but a popular concept, argued Schuchardt, which linguists had
adopted but were unable to adapt to their scientific goals. He then revealed its
problematic nature in a surprisingly accurate and concise fashion:
On the one hand, it is something relative, with dialect next to itself and language
above itself; on the other hand, it is something complex, lacking a required
coherence of its parts and lacking a fixed delimitation, so not an individual
object, no organism, as it has so often been conceived.¹³
The gist of what followed boils down to this: no clear borders between dialects
and languages could be drawn, and linguists who had done so relied on only a
selection of mainly phonetic features. And if there were any borders, these were
¹⁰ See the accompanying preface in Schuchardt (1900: 1*) as well as Schuchardt (1928: 193). On the
contact phenomena see in particular Schuchardt (1900: 10–11).
¹¹ Schuchardt (1900: 12): ‘nur geographische Kollektivausdrücke’. Cf. Schuchardt (1928: 119): ‘die
Mundarten sind uns kaum mehr als geographische Begriffe’.
¹² Schuchardt (1922: 10): ‘das allgemeine Mundartenproblem’.
¹³ Schuchardt (1922: 10): ‘Einerseits etwas Relatives, mit Mundart neben sich und Sprache über sich;
anderseits etwas Komplexes, ohne notwendigen Zusammenhang der Teile und ohne feste
Umgrenzung, also kein Individuum, kein Organismus, wie man es so oft aufgefaßt hat’.
. ,  239
arbitrary and due to a long-term interruption of contact between different speech

communities. In deploying this argument, Schuchardt once again referred to his
1870 lecture on Romance dialects. Linguists, he noticed, seemingly could not do
without dialects, much as geographers could not do without their precious system
of coordinates. And they proceeded wrongly, since they first selected a number of
features out of all linguistic facts and then claimed that these were responsible for
constituting a dialect. On this limited basis, they construed the linguistic entity of
dialect. As a result, it was no more than a ‘mental object’ (Gedankending), to
which they had a priori assigned a number of essential properties. They started, in
other words, from a fictitious concept and a limited set of features rather than
from linguistic facts when studying language variation.
Linguists were therefore, Schuchardt ingenuously judged, mere shibbolethists
(Schibbolethiker). Just as the biblical Gileadites had used the Hebrew word
shibbólet to identify and kill the enemy Ephraimites, who allegedly pronounced
it as sibbólet, linguists relied on a few features only to identify distinct dialects, thus
paralyzing research into language diversity. Looking at other features, Schuchardt
implied, might have led them to different, equally fallacious conclusions. If there
was any secure ground at all to introduce dividing lines into the continuum of
language, it should be sought in the rhythm and musicality of a variety, which
reflected a tribe’s individuality much better than phonetic features did.¹⁴ Yet this
was more a cautious proposal by Schuchardt than an incontestable maxim, and he
felt forced to conclude that no clear borders could be drawn between dialects, only
between groups of linguistic facts. The latter should therefore be the main
ingredient of language atlases, for individual features reflected the natural dyna-
mism of language and its inherent mixedness much better than fictitiously static
entities such as dialects and languages.
After all, language was part of nature and fluctuated with it—or as Schuchardt
put it with the famous saying of the Ancient Greek philosopher Heraclitus (fl. c.
500 ): in language, too, ‘everything flows’.¹⁵ For this reason, Schuchardt argued
in his review of Ferdinand de Saussure’s Cours de linguistique générale (1916) that
it was pointless to make a distinction between general and language-specific
linguistics, since ‘there is only one, which at one time explores here, at another
time there, which scrutizines at one time this matter and at another time that’.¹⁶
‘Language’, Schuchardt continued, ‘constitutes a whole, a continuum’.¹⁷ At least,
that was how it should be conceived, since there were transitions rather than fixed
dividing walls in it.
¹⁴ Schuchardt (1922: 11), speaking of ‘die Eigenart des Stammes’.

¹⁵ Schuchardt (1922: 15). The original Greek saying is ‘pánta rheî’ (‘πάντα ῥεῖ’).
¹⁶ Schuchardt (1928: 318): ‘es gibt eine einzige, die bald hier, bald dort schürft, bald diesen, bald
jenen Stoff durchleuchtet’.
¹⁷ ‘Die Sprache bildet eine Einheit, ein Kontinuum.’
240   
18.5 Useless abstractions?
In Schuchardt, the modern scepticism of the conceptual pair culminated. It had

been ever-growing in the latter half of the nineteenth century, but Schuchardt
carried the matter to an extreme and believed he had revealed the true nature of
the concepts language and dialect. They were useless abstractions, and linguists
should follow his lead and focus on the facts rather than on mental objects
grounded in a priori assumptions. Schuchardt, in other words, was questioning
one of the most fundamental premises of all language studies up to then: the idea
that ‘languages or dialects were self-contained units easily identifiable’ (Morpurgo
Davies 1998: 285). Even if linguists such as August Friedrich Pott (1856: 220) had
expressed suspicions about the conceptual pair some decades before Schuchardt,
they never considered the full consequences of the problem and never adopted an
attitude as radical as Schuchardt’s.
Several of Schuchardt’s contemporaries, too, realized the problematic character
of the language/dialect distinction, but most drew less extreme conclusions. The
French comparative linguist Victor Henry (1850–1907), for instance, treated the
matter in thesis/antithesis format in his Linguistic antinomies of 1896, an intri-
guing work with relatively poor distribution.¹⁸ Henry put forward the following
thesis: ‘The categories of language, languages and dialects, even of simple words, if
examined closely, are only abstractions with no external reality’.¹⁹ The accom-
panying antithesis read: ‘There exists a science of language, which takes as its
object of study the phenomena of the life of language, that is, the life of languages
and the life of words’. In other words, linguistics studied language, but language
was an abstract concept which did not correspond to anything in actual reality.
This mismatch was problematic, Henry granted, especially since a ‘science does
not usually debut by declaring itself without an object’ (Henry in Joseph 1996: 125;
cf. Morpurgo Davies 1998: 290).
Although repeatedly emphasizing that concepts such as language and dialect
were mere abstractions deprived of reality and suggesting that perhaps ‘the
objective notion of the infinity of speakers’ should replace ‘the empty abstraction’
that was language, Henry felt forced to comply with tradition based on practical
considerations:
But, after all, what matters are not words, but clear ideas. And if words can clarify
ideas, it is on the condition that they not shock established traditions too much.
So let us continue—we have no choice—to speak of language, even though we
know that there is no language, but only people who speak. Let us continue to
¹⁸ See Joseph (1996); Morpurgo Davies (1998: 288). Noordegraaf (2014: 525) nuances the poor
distribution of the work.
¹⁹ I quote from Joseph’s (1996: 124–37, here 124) translation of the first chapter of Henry’s work.
.   241
speak of language families, of particular languages, of dialects, sub-dialects,

patois, correct or incorrect pronunciation, provided that we understand in
every case, beneath each of these words, a single latent meaning, the same for
everyone: namely, in each of the irreducible units of language that science has
provisionally established, an indefinite series of variations which get attenuated
imperceptibly as one descends from the race to the nation, the province, the
canton, the family and finally the individual—or rather which are parts of
individual and initially imperceptible nuances of pronunciation and expression,
and which are aggravated by circumstances until they end up at dialectal scission
or even at linguistic isolation. (Henry in Joseph 1996: 128)
In short, to Henry it seemed that the concepts maintained their practical use as
analytical linguistic tools, a conclusion with which Schuchardt no doubt would
have disagreed.
Henry’s views on the concepts of language and especially dialect were, however,
not shared by a number of his compatriots who pioneered Gallo-Romance
dialectology. In the French dialectological tradition, a true Schuchardtian spirit
came to dominate, personified by linguists of the calibre of Gaston Paris
(1839–1903), Paul Meyer (1840–1917), and Jules Gilliéron (1854–1926). These
scholars did not believe in the existence of dialects, only linguistic features, a bold
claim at the time, later embodied by Gilliéron’s monumental Linguistic atlas of
France (1902–10).²⁰ The evidence did not show any obvious dialect divisions but
rather fluctuating isoglosses, these French dialectologists claimed, as they lacked a
holistic view on linguistic diversity. Notably, Paris and Meyer, both of whom were
familiar with German methods of language study, were involved in a long-running
controversy over dialect lines between French (langue d’oïl) and Occitan (langue
d’oc) with the Montpellier school of dialectologists, who did work with a concept
of self-contained dialects, most notably Octavien Bringuier (1820–75) and Charles
de Tourtoulon (1836–1913).²¹
The mixed nature of dialectal diversity was acknowledged also by linguists
outside the dialectological tradition, including Michel Bréal (1832–1915), famous
for his foundational essay on semantics, in which he criticized the illusion of pure
and unmixed dialects (Bréal 1897: 3–4). The French scepticism constituted a
reaction against the neogrammarian mechanistic conception of language, which
presupposed regular development and discrete entities (cf. Kretzschmar in Boberg
et al. 2018: 61). Hans Goebl (1982; 2003) has fittingly labelled the French approach
typophobia, a fear of assuming abstract linguistic entities with clear boundaries, as
²⁰ See in particular Goebl (1982; 2003) and Goebl in Boberg et al. (2018: 124–5). Cf. also Desmet
(1996: 459); Engler (2000); Desmet et al. (2002); Joseph (2012: 462).
²¹ For an excellent synthesis of the frictions between both groups of French dialectologists as
‘communities of practice’ see most recently Klippi (2020: esp. 114–16 on the dialect concept). See
also Hoyt (2006).
242   
opposed to the typophilia prevailing in German scholarship and with Graziadio

Isaia Ascoli (1829–1907) also in Italy. Ascoli (1876) actively reacted against the
views of Paul Meyer and Gaston Paris, with whom he had a dispute over Franco-
Provençal dialect lines. In Ascoli’s (1876: 387) view, there were, in fact, separate
dialect entities, marked by the ‘simultaneous presence’ or ‘particular combination’
of certain linguistic features.²² He thus adopted a more holistic perspective on
regional language diversity than the Parisian school of dialectologists.
The opposition between French (Parisian) and German–Italian dialectology
also shaped later dialectological traditions. The former tended to focus on words
and individual linguistic features, involving social factors in their analysis, whereas
the latter tried to discover dialect areas in a more strictly regional sense. Scholars
could change their minds as to which approach was preferable. A notable case is
the Swiss linguist Louis Gauchat (1866–1942), who, trained in the Parisian
dialectological tradition, was typophobic at first, before siding with the typophilic
camp, developing the concept of core landscape (Kernlandschaft) to chart dialectal
diversity in centres and peripheries (e.g. Lang 1982: 206–7; Desmet et al. 2002:
33–6). Reconciliatory approaches also existed, such as the one adopted by the
Austrian-American dialectologist Hans Kurath (1891–1992; see Kretzschmar in
Boberg et al. 2018: 63).
18.6 Conclusion
Even though there previously had been suspicion of the language/dialect distinc-
tion, most nineteenth-century and early twentieth-century linguists kept on using
it as a self-evident device in a very flexible way, priming one interpretation or
another as it suited them. It was Hugo Schuchardt who fully exploited its
problematic nature for the first time by tearing down the walls built between
languages and the fences erected within them between dialects ever since the
sixteenth century. Schuchardt did so in incisive comments scattered throughout
his diversified oeuvre, as he did not produce a grand theoretical work. He richly
deserves Anna Morpurgo Davies’s (1998: 288) well-chosen label of ‘iconoclast’,
since he radically questioned a number of basic assumptions in linguistics,
including the conceptual pair. Schuchardt was, however, most certainly not an
isolated thinker but the chief exponent of a more general tendency, as the trends in
French dialectology suggest.
Linguists increasingly questioned fundamental concepts such as language and
dialect, which led to the definitive abandonment of the influential organicistic
conception of language, advanced and popularized by August Schleicher. The
²² ‘ma il distintivo necessario del determinato tipo sta appunto nella simultanea presenza o nella
particolar combinazione di quei caratteri.’
.  243
question of ‘What is (a) language?’ was now wide open, yet many linguists chose
to avoid and even abandon abstract concepts such as language and dialect and to
focus instead on concrete linguistic facts and features—very often words—
especially as the results of the new subfield of dialect geography confirmed the
problematic nature of such abstract concepts (Morpurgo Davies 1998: 289–90;
Boberg et al. 2018: 7–8). The linguistic facts, many linguists realized, did not allow
anybody to presuppose clearly distinct and self-contained units such as languages
and dialects. Language, then, was no longer countable. In mainstream linguistics,
however, the conclusions of Schuchardt and the French dialectologists came to be
overshadowed by the grand theories of the twentieth century.
19
From Saussure to 1954
Structuralism and the language/dialect
distinction
The closer I come to the present, the more difficult it becomes to provide a
satisfying historical account, not only because of the enormous wealth of source
material and the vast diversity of ideas and increasingly complicated approaches
encountered in it, but also because the distance from the sources shrinks and
secondary literature becomes much more limited.¹ For this reason, it is simply
impossible to offer a detailed outline of the history of the conceptual pair in post-
1900 linguistics, which lies outside the scope of the present overview and deserves
a study in its own right. I will limit myself to distinguishing a number of notable
tendencies in the last 120 years within mainstream linguistics and the ways in
which they relate to the previous history of the language/dialect distinction and
are anchored in it. I will do so on the basis of the works of key linguists like
Ferdinand de Saussure, as well as of recent handbooks in the field of linguistics
and, especially, dialectology and language variation. The latter provide useful
overviews of the diverging approaches to dialectal variation common in the last
century and a half and thus constitute fine embodiments of the various attitudes
modern linguists have adopted vis-à-vis the conceptual pair (see e.g. Boberg et al.
2018). A study of these text types will be my basis for providing a first rough
sketch of the latest episode in the history of the language/dialect distinction, from
which later studies can hopefully benefit.
What was the place of the conceptual pair in the main currents of post-1900
linguistics, including most notably structural, generative, and sociolinguistic
approaches to the phenomenon of language? In an overview that is partly chrono-
logical and partly thematic, I will argue that this question can be answered in
terms of a love-hate relationship, since linguists’ attitudes have oscillated between
two extremes: a deep-seated belief in the use of the language/dialect distinction
and its downright dismissal.
¹ For an indication of the wealth of source material see e.g. the anthology in Göschel et al. (1976) and
the papers in Göschel et al. (1980) as well as the chapters and references in Boberg et al. (2018). Cf. e.g.
also Shapiro and Schiffman (1981: Chapter 2).
DOI: 10.1093/oso/9780198845713.001.0001
.   ’   245
19.1 In Saussure’s class
The beginning of the twentieth century witnessed a major turn in language

studies, personified by no one better than the Swiss linguist Ferdinand de
Saussure (1857–1913; see Joseph 2012 for an extensive biography). Whereas
during his lifetime Saussure published mainly on classical topics from Indo-
European historical-comparative linguistics, including a ground-breaking disser-
tation on the Proto-Indo-European vowel system, he adopted a very different
perspective on language in some of his courses at the University of Geneva,
published only posthumously in 1916 by his students Charles Bally and Albert
Sechehaye (see Saussure 1995 [1916]).
Saussure famously introduced a distinction between langue, language as a
socially shared and structured system, shaped by the relationships between
signs, and parole, actual usage of language in context. Trained as a philologist,
he reflected extensively on both of these aspects of language. Over time, Saussure
became the symbol of the shift in mainstream linguistics from traditional
historical-comparative linguistics to structural linguistics, despite his interest in
both approaches. This shift involved a major change in perspective from
diachrony to synchrony, even if traditional forms of linguistics, historical-
comparative grammar as well as dialectology, persisted, albeit now relegated to
the margin of mainstream research.
Language as a system abstracted from time and place became the main object of
linguists for about half a century, first in structuralism, then in generative gram-
mar as developed by Noam Chomsky from the 1950s onwards. Since linguists
operating within these strands were primarily interested in language as a self-
contained system, linguistic diversity was not high on their research agenda (see
e.g. Bierbach 1978: 75). Instead, the starting point was the ideal speaker, often
coinciding with the linguist himself, and this ideal speaker did not usually speak
dialect. It is telling that no full-fledged structural and generative dialectologies
have developed, even though there have been some attempts in this direction, with
varying success (see Boberg et al. 2018: Chapters 4–5; the next section, 19.2;
Chapter 21, Section 21.1). The comparative philologist’s books and the dialect-
ologist’s questionnaires, one might say, ceded to the armchair, from where the
linguist could comfortably think about the underlying system of language. What
were the consequences of this turn in mainstream linguistics for ideas about the
conceptual pair?
The consequences seem to have been limited, at least initially. A pivotal figure,
Saussure took a rather traditional, although not unambiguous, view on the
language/dialect distinction. Saussure agreed with his French colleagues Gaston
Paris and Paul Meyer that it was impossible to draw ‘clear boundaries’ between
dialects, especially since ‘languages are in endless flux’. In his course on
246    
geographical linguistics, he even actively taught Meyer’s pointed observation that

there were no such things as dialects but only dialect features.² The fuzziness of
linguistic borders forced him to think about the conceptual pair and its validity, a
topic he broached in his third general linguistics course, taught in the winter
semester of 1910–11. One of the students attending this course was Emile
Constantin, whose extensive notes have been published in 2005 and which
I take as the basis of my discussion.
Constantin’s notes suggest that Saussure treated the language/dialect distinction,
and especially the concept of dialect, at some length in his classes. The occasion for
the Swiss professor to do so was a passing usage of the French term dialecte, which he
mentioned while pointing out that linguistic difference was a matter of gradation.
Latin and Greek were more closely related to each other than they were to Sanskrit,
Saussure remarked. Even closer kinship existed among dialects, at which point he
observed:
But in pronouncing this word, we immediately add that one should not attach
any absolute idea to the term dialect in relation to language. There is no precise
point where the designation dialect is reached instead of the designation
language.³
Some had proposed mutual intelligibility as a criterion, Saussure was well aware,
but he does not seem to have been convinced of its utility. This phenomenon, too,
was a question of gradation, and cutting the continuum in half would be an
arbitrary act (Saussure 1995 [1916]: 278–9). There is an underlying contradiction
here. Saussure on the one hand refrained from granting existence to dialects—
there were only dialectal features—but he supposed on the other hand that
abstract entities such as languages and dialects did, in fact, exist, even if it was
difficult to say when a dialect started being a language. The Swiss linguist was
aware of this discrepancy and believed that ‘in practice, one should preserve the
term dialect’, yet that this should occur under certain conditions.⁴ Linguists should
either ‘agree that one single feature suffices to characterize a dialect’ or if they
considered all features, they should ‘limit themselves to a single point on the map
and speak of the dialect of some town’.⁵ And even then, dialects should be
² Joseph (2012: 379, 462–3). See Constantin and Saussure (2005: 117–26). Cf. Chapter 18,
Section 18.5.
³ Constantin and Saussure (2005: 100): ‘Mais en prononçant ce mot, nous ajoutons tout de suite
qu’il ne faut attacher aucune idée absolue au terme de « dialecte » par rapport à celui de langue. Il n’y a
aucun point précis où intervienne le nom de dialecte au lieu du nom de langue’.
⁴ Constantin and Saussure (2005: 123): ‘Dans la pratique, il faut conserver le terme de dialecte’.
⁵ Constantin and Saussure (2005: 121): ‘Pour qu’il y ait un dialecte : ou bien il faut 1 ) convenir
qu’un seul caractère suffit pour caractériser un dialecte, 2 ) ou bien, si l’on prend tous les caractères, il
faut s’enfermer sur un seul point de la carte et parler du dialecte de tel village’. Cf. Saussure
(1995 [1916]: 276).
.     247
conceived as open rather than closed entities, which took part in broader waves of
linguistic innovation, Saussure claimed with clear reference to the wave model
(Constantin and Saussure 2005: 125–6; see Chapter 18, Section 18.1). With these
conditions in mind, one could continue speaking of dialects; otherwise one was
forced to conclude that there were no such things (see also Kretzschmar 2009:
50–3, 55–8; Joseph 2012: 571, 731 n.32).
Saussure attributed other values to the language/dialect distinction, however.
Whereas he basically rejected the conceptual pair as an adequate means to map
out regional linguistic diversity because of its gradational and uneven nature, he
maintained it in two other traditional senses: the language-historical interpret-
ation of language mother/daughter dialect and in the opposition standard lan-
guage/dialect.⁶ With regard to the latter sense, he observed that ‘in a natural
language, there are only dialects; a language left to itself is destined to indefinite
fractioning’.⁷ From these dialects, that of the most civilized and powerful province
was selected to become the literary and general language and was usually influ-
enced by other provinces in due course (Constantin and Saussure 2005: 102–3;
cf. Saussure 1995 [1916]: 267–8).
Saussure thus recognized that it was impossible to find a purely linguistic
criterion to distinguish languages from dialects—if there were any dialects at
all—and at the same time that the conceptual distinction was shaped by extra-
linguistic factors, most notably culture and politics. In conclusion, the Genevan
professor was fairly unoriginal when it came to his conception of the language/
dialect pair. He followed trends in French dialectology, with which he was well-
acquainted, since he had spent several years in Paris as a lecturer at the École des
Hautes Études (Joseph 2012: Part III).
19.2 Towards a structural dialectology:

Uriel Weinreich’s diasystem
In Saussure’s wake, the focus was for a long time on langue, the language system,
and usually not on linguistic diversity, relegated to the parole domain (cf. Gordon
in Boberg et al. 2018: 75). His colleagues and students limited themselves to
noticing the gradual nature of linguistic variation (e.g. Sechehaye 1908: 251), if
they touched upon the matter at all. Indeed, structural linguists of various strands
seem to initially have had the tendency to insist on the relative nature of the
conceptual pair, while at the same time continuing its usage, most often without
⁶ For an example of a reference to an Indo-European dialect see Constantin and Saussure

(2005: 170).
⁷ Constantin and Saussure (2005: 102): ‘Dans une langue naturelle, il n’y a que des dialectes : une
langue laissée à elle-même est vouée au fractionnement indéfini’.
248    
clearly defining it. An exception was the American structuralist Leonard

Bloomfield (1887–1949), who proposed in Language, his 1933 linguistics text-
book, to expand the conceptual pair in order to overcome its binary nature.
Bloomfield (1933: 52, 54) did so by integrating it into a five-layer division of
literary standard, colloquial standard, provincial standard, substandard, and local
dialect. This cutting up of the variation continuum was most likely as arbitrary as
the distinction between language and dialect. Although more nuanced, it was
based on a mixture of exclusively language-external criteria, principally geo-
graphic and sociocultural parameters.
Ambiguous attitudes towards the language/dialect opposition were common
among linguists active in the first decades of the twentieth century in general, also
outside the sphere of structural linguistics strictly speaking. For instance, dialect
was an ‘intangible notion’ according to the French linguist Antoine Meillet
(1866–1936), Saussure’s successor in Paris.⁸ By consequence, the concept was
not very useful to a comparative linguist, who should look for a more clearly
defined research object. However, Meillet’s suspicion did not keep him from
eagerly using the term dialecte himself throughout his rich oeuvre, mainly in its
language-historical sense, in phrases such as ‘the Indo-European dialects’
(cf. Meillet 1908: 4–5; 1925: 53–9).
A similarly ambiguous attitude was adopted by the German-born scholar
Edward Sapir (1884–1939) in his textbook Language (Sapir 1921: 159–64).
Inspired by his teacher Franz Boas (1858–1942), Sapir approached language
principally from an anthropological point of view in the United States and
paved the way for the later American structuralist tradition, closely associated
with the figure of Leonard Bloomfield. Sapir did, however, contribute to the
discussion about the concept of dialect by taking a sociological view on it, in
which he did not presuppose any strict language/dialect opposition. In his entry
for ‘dialect’ in the Encyclopaedia of the Social Sciences, he remarked that a ‘group
of dialects is merely the socialized form of the universal tendency to individual
variation in speech’ (Sapir 1931: 123), thus linking dialect to idiolect rather than
language (cf. Göschel et al. 1976: –).
Given the ambiguity of the attitudes which early twentieth-century linguists
adopted towards the conceptual pair, the reader often has to infer their concep-
tions from their terminological usage. To stick with the example of Sapir, it is clear
from his manual Language that he entertained traditional conceptions of the
language/dialect distinction. Although noticing that the distinction is relative
and arbitrary and nothing more than a snapshot in time, Sapir (1921: 163–4)
seems to have assumed that two related dialects became different languages as
soon as they were no longer mutually intelligible, even though he did not explicitly
⁸ See Meillet (1924: 15): ‘Cette notion est fuyante’.

.     249
posit this criterion. Central to his conception of the distinction was the traditional
language-historical interpretation. A language developed into dialects, which in
turn became languages again in due time. In a footnote, Sapir (1921: 164 n.6)
dismissed the interpretation of dialect ‘in contrast to an accepted literary norm’ as
irrelevant to his linguistic research.
Linguists continued to harbour only superficial suspicions about the conceptual
pair well into the 1950s. Symptomatic of this state of affairs is that they did not
engage in any direct debate. This lack of interaction was, as I have argued in
previous chapters, a widespread phenomenon in the history of the language/
dialect opposition, in which there were only a few flashes of animated discussion,
usually exhibiting the same pattern: one or a few scholars reacted against common
practices, like Johann Heinrich Hottinger in the seventeenth century (Chapter 10,
Section 10.2) and Hugo Schuchardt in the decades around 1900 (Chapter 18). In
the 1950s, structural linguists were challenged from within to involve regional
variation in their study of language, thus sparking off a more direct and lively
debate. The possibility of a structural dialectology was finally examined, an
endeavour spurred not in the least by the Polish-American linguist Uriel
Weinreich (1926–67; see Malkiel 1967), son of Max, the Yiddish specialist who,
it will be remembered, first recorded the quip ‘A language is a dialect with an army
and navy’ (see Chapter 1).
Weinreich the younger studied linguistics at Columbia University with André
Martinet (see the next section, 19.3). For his PhD dissertation (1951) and the
resulting foundation-laying 1953 book Languages in Contact, he conducted field-
work in linguistically diverse Switzerland, where he also came into contact with
the work of local dialectologists, which he valued highly (cf. Weinreich 1954: 397
n.18). His structuralist formation as well as his interest in language-contact
phenomena and linguistic diversity led him to ask himself: ‘Is a structural dialect-
ology possible?’ In an influential and foundational paper, bearing this question as
its title and published in Word, a journal of which he was co-editor, Weinreich
(1954) answered it favourably by providing some schematized examples; he took
them mainly from phonology, the domain where structuralism achieved most of
its successes. He did, however, emphasize that this linguistic approach should
occur in conjunction with an externalist form of dialectology, which sought to
explain linguistic differences by correlating them with cultural areas and borders
(Sprachlandschaft). This form of dialectology was promoted especially by Swiss
linguists in reaction to the idolatry of isoglosses in earlier scholarship.
Weinreich wanted, in sum, to bridge the gap between structuralism and
dialectology. What was the place of the conceptual pair in his attempts at
reconciliation? Apparently, it did not have one, since he renounced the common
usage of the term dialect and its many connotations, because it was ‘non-technical’
and defined by language-external, non-structural attributes of time, space, and
prestige (Weinreich 1954: 399). He, too, in other words, probably believed in the
250    
validity of the witticism associated with his father. Weinreich (1954: 397 n.17) also
found it confusing that the term dialect was used in phrases such as ‘Indo-
European dialects’ and ‘literary dialects’. The eternal problem of dividing dialect
continua into separate dialects likewise fostered his suspicion, since it could not be
solved by simply unveiling isogloss bundles on a map. Weinreich (1954: 397)
concluded:
It is evident that no unambiguous concept of dialect could emerge even from this
optimistic methodology any more than a society can be exhaustively and
uniquely divided into “groups”.
Yet if structuralists and dialectologists were to join forces, Weinreich (1954: 399)
did not rule out ‘the formulation of a technical concept of “dialect” as a variety or
diasystem with certain explicit defining features’.
As long as the term dialect did not receive a technical interpretation,
Weinreich preferred to operate with the concept of diasystem in his proposed
structural dialectology. In Saussure’s wake, he explained, linguists have thus far
concentrated on language as a uniform system, an illusion upheld by their usual
object of study: standardized rather than folk language. To integrate the latter
into structuralist theory, Weinreich suggested conceiving of a diasystem as
follows. For structuralists, language is a ‘unique and closed system whose mem-
bers are defined by opposition to each other and by their functions with respect
to each other, not by anything outside of the system’.⁹ Structuralists should now
proceed to
constructing systems of a higher level out of the discrete and homogeneous

systems that are derived from description and that represent each a unique
formal organization of the substance of expression and content . . . A “diasystem”
can be constructed by the linguistic analyst out of any two systems which have
partial similarities (it is these similarities which make it something different from
the mere sum of two systems). But this does not mean that it is always a scientist’s
construction only: a “diasystem” is experienced in a very real way by bilingual
(including “bidialectal”) speakers and corresponds to what students of language
contact have called “merged system”.
Dialectal variation could, in other words, be regarded as an ‘aggregate of

systems’ with different layers, which Weinreich preferred to term varieties rather
than dialects. Apparently, variety was in his eyes a more neutral term than dialect
with its dubious connotations. His proposal has proved to be very successful, since
variety is still widely used as a passe-partout term by linguists studying variation,
⁹ All quotes in this paragraph are from Weinreich (1954: 388–90).

.     251
even though some scholars have criticized its excessively noncommittal nature
(see e.g. Macaulay 2010: 62–3).
For Uriel Weinreich, in short, dialectology was the ‘study of diasystems’ and the
varieties that constituted those diasystems, apparently now presupposing that they
were easy to distinguish from each other. The term dialect, he judged, should be
abandoned, at least for now, until a scientific definition was reached. Old habits
die hard, however, and traces of the age-old language/dialect distinction remained
latently present in his work, including his article on structural dialectology.
Weinreich referred, for instance, to bilingual and bidialectal speakers, people
mastering two languages or two dialects; this terminology suggests how deeply
rooted the conceptual pair had become in the minds of linguists, even those
renouncing the term dialect as non-scientific.
19.3 Redefining the conceptual pair: Martinet and Polák
Around the same time as Weinreich was developing his thoughts on structural
dialectology, some of his colleagues attempted to remedy the non-technicality of
dialect and to provide a sound basis for distinguishing language from dialect.
These included, most notably, the French linguist André Martinet (1908–99) and
the Czech structuralist Václav Polák (1912–81), both of whom Weinreich cited in
his structural dialectology paper, although admitting that the latter’s article had
arrived too late to be fully acknowledged.
Weinreich’s teacher Martinet wrote an article entitled ‘Dialect’, published in the
journal Romance Philology in 1954, in which he problematized the concept named
in its title, because he had noticed how loosely it was being used by contemporary
dialectologists. He complained that Sever Pop (1901–61) in his monumental La
dialectologie of 1950 had not touched upon so central a concept as dialect, and that
dialectologists from the French school, like Gilliéron, had not addressed this issue
either, because they did not believe in the existence of dialects and patois, only in
isoglosses. This overall neglect compelled him to reflect on the matter himself.
Martinet set out by deploring the proliferation of poorly defined terms to talk
about various forms of language, including the terms language and dialect. The
latter was, for instance, used to refer to allegedly deviant forms of a standard
language as well as to legitimate forms of a language. Linguists had, moreover,
confounded speech forms derived from a common ancestor with variations on a
later developed standard.
Martinet’s solution was to formulate more objective interpretations. Genetic
kinship, he maintained, should be the primary criterion in deciding whether a
speech form was a language or a dialect:
As a rule, genetic kinship would seem to be the decisive criterion: Corsican rates
as an Italian dialect because it is genetically close to Tuscan, the dialect on which
252    
the Italian “koinê” is based. Corsican is not a French dialect although its speakers
have recourse to French as a standard. (Martinet 1954: 7)
The existence of a koine, believed Martinet, could be a secondary criterion for

using the term language. He was, however, fully aware that sociopolitical factors
often interfered with this objective criterion of genetic kinship, even among
linguists. Basque, for instance, was usually regarded as a distinct language because
of the genetic criterion, whereas Catalan was frequently claimed to be a Spanish
dialect for political reasons:
A form of speech of indefinite status will be denied the title of “language” if it can
somehow be labeled a dialect of a national language that enjoys the sanction of a
special color on the map.
This mixing up of criteria, Martinet (1954: 8) realized, made the linguist look bad,
who ‘tends to be just as hesitant as the layman, because actually both use the same
terms, and the linguist has simply never taken the trouble to redefine them
scientifically’.
The linguist should accept responsibility for the terminology of his discipline,
according to Martinet, who challenged his colleagues to think about the language/
dialect distinction rather than simply dismiss or relativize it. This redefining of
existing terms was, he claimed, also more economical than inventing new ones. In
the case of language and dialect, redefinition was all the more appropriate, since
the concepts were already in the beloved format of structuralists, the opposition of
linguistic units, even though he did grant that language variation was not a simple
binary matter, but a gradational phenomenon. For this reason, he added an extra
layer to the conceptual pair, that of patois, referring to the speech form of a small
community, typically a village, that had lost its overarching provincial dialect and
in consequence greatly differed from the standard language of the country. He
further proposed the more neutral term vernacular as a near-synonym of patois,
which carried negative connotations.
Martinet’s tripartite division was not as new as he wanted his readers to believe.
His master Antoine Meillet, for instance, had already suggested a similar hierarchy
in a work on the comparative study of Indo-European tongues (Meillet 1903: 31).
In the early modern period, too, linguistic diversity had already been charted in
terms of language, dialect, and subdialect. Martinet was, however, innovative in
teasing out the historical conditions in which dialects and patois emerged. He
reacted against the exclusive focus on dialect divergence, central to conceptions of
regional variation since the sixteenth century: from a language into dialects, which
became in time separate languages, which itself developed dialects again. This
supposed standard evolution was too simple a picture of sociolinguistic reality,
since there were cases of dialect convergence, too, Martinet pointed out.
.     253
Although genuinely concerned over the language/dialect distinction and how to

redefine it scientifically, Martinet proposed a rather poor conceptualization him-
self, since he did not further discuss how his ‘decisive criterion’ of genetic kinship
should work exactly. It does seem, however, that he had different degrees of
kinship in mind, thus perhaps presuming a variation on the Aristotelian criterion
(see Chapter 8, Section 8.1). Yet Martinet did not suggest any cut-off point or
other method to convincingly differentiate a dialect from a language.
Also in 1954, the Czech structuralist Václav Polák, active in the Prague linguis-
tic circle, published an article wholly devoted to the distinction between language
and dialect (Polák 1954). In this contribution, which appeared in the Leuven-
based journal Orbis, founded by Sever Pop, Polák attempted to establish a purely
linguistic foundation for the conceptual pair, as if he had anticipated Martinet’s
call to responsibility. He, too, considered it indispensable that linguists defined
their core concepts; he, too, emphasized the importance of linguistic convergence,
especially for dialect studies. Finding fault with earlier linguists such as Graziadio
Ascoli, Hermann Paul, Hugo Schuchardt, and Antoine Meillet, whom he criti-
cized for their usage of extra-linguistic criteria in defining the conceptual pair,
Polák set out to design a strictly linguistic approach to the matter. In typically
structuralist fashion, he interpreted dialect as part of the system of a language—a
kind of subsystem—which was why it should be understood in linguistic terms.
For Polák, dialects resulted from the confrontation of the linguistic system with
neighbouring systems, framing it within the concept of Sprachbund, or union des
langues in his own French. Egyptian and Coptic, Polák’s beloved example,
belonged to the same linguistic system, but through contact with other Hamitic
and Bantu tongues Coptic had acquired certain structural characteristics which
Egyptian did not have, but which did not cause language-level differences. The
structural frame of a language was, in other words, timeless, Polák believed; he
used the terms panchronique and achronique himself (Polák 1954: 96). It was only
if one perceived distinct morphosyntactic systems that one could speak of differ-
ent languages. Variation on the level of phonetics and lexico-semantics only
generated different dialects or stylistic variants.
Polák, in other words, refashioned the conceptual pair in distinctly structural
terms in a way vaguely reminding of the early modern Aristotelian criterion.
Related dialects exhibited superficial differences belonging to the phonetic
and lexical level, whereas distinct languages had separate morphosyntactic sys-
tems, the dominant, most essential, and most stable layer of a language. Polák’s
(1954: 93) conscious choice to consider diversity from an exclusively language-
internal perspective implied excluding any social factors whatsoever. It was such
isolationist approaches to language that the social turn in language studies aimed
to counter in the following decade (see Chapter 21, Section 21.2). Polák did,
however, make a distinction between horizontal and vertical varieties of a lan-
guage system. The former were the dialects, coinciding with regional subsystems
254    
of a language, the latter were stylistic variations on the language system such as
jargon, slang, and literary language, situated mainly on the level of the lexicon.
A purely linguistic, system-based interpretation of the conceptual pair, argued
Polák, had practical use, since it was an objective means whereby to distinguish
independent languages from each other, especially in cases where there were not
yet boundaries linked to literary languages, as in the Caucasus, or where linguists
had been using the terms language and dialect sloppily, as in the case of the Slavic
family of tongues. Polák believed that the latter—with the exception of
Bulgarian—should be conceived of as dialects of one and the same language
because they shared in their core the same morphosyntactic system. Linguists,
however, usually considered them to be distinct languages, but this view was
shaped by extra-linguistic criteria related to politics, culture, and civilization.
They were, one might paraphrase, exalting these Slavic dialects to language status
because they had an army, navy, and literature. In his exceptional attempt at tying
the language/dialect distinction to specific linguistic properties and accommodat-
ing it to structuralist conceptions of language, Polák failed to address the eternal
problem of how to distinguish between dialects. His suggestion that they were
‘different stages of the continuous variation of a specific linguistic structure’
indicates, however, that he considered it possible to establish some kind of borders
between dialects.¹⁰
19.4 Conclusion
The year 1954 was a key turning point in the history of the conceptual pair, since
structural linguists from different backgrounds suddenly tried to define it in more
objective terms. They did so independently from each other, proposing diverging
solutions, and they were motivated to do so by the loose and confused usage of the
terms language and especially dialect current in their discipline and among the
general public. Especially Weinreich’s paper on structural dialectology met with
considerable success. His concept of diasystem, which one might regard as a
structuralist reinterpretation of the language/dialect distinction, as well as his
annotation method for dialect features became widely known and applied but
also criticized (see e.g. Gordon in Boberg et al. 2018: 83–4).
Some structural linguists, however, did not consider the flexible usage of the
terms language and dialect a problem, but an asset. Most notably, according to the
American scholar Charles F. Hockett (1916–2000) in his 1958 handbook A Course
in Modern Linguistics, ‘[t]he relative looseness of the two terms is a merit, not a
defect, for one can add as many precisely delimited technical terms as one needs,
¹⁰ Polák (1954: 98): ‘étapes diverses de la variation continue de la structure de langue en question’.
.  255
based on various criteria of similarity between idiolects’.¹¹ Typically for the

American structuralist tradition, Hockett stipulated the idiolect, the speech form
particular to an individual, as the lowest level of linguistic diversity, which
constituted the basis for research and from which higher-level abstractions were
to be made. Both language and dialect were collections of similar idiolects, he
suggested, with the difference that ‘the degree of similarity of the idiolects in a
single dialect is presumed to be greater than that of all the idiolects in the
language’. Despite claiming that the looseness of the terms language and dialect
was an asset and fostered a flexible approach to linguistic diversity, Hockett did
propose a specific criterion ‘stem[ming] from an everyday assumption about
language’: mutual intelligibility.
¹¹ All citations in this paragraph are from Hockett (1958: 322).

20
Mutual intelligibility
The number one criterion?
Charles Hockett believed that the intuition among laypeople that speakers of
related dialects could understand one another, but those of distinct languages
could not, required a more objective formulation. Aware that mutual intelligibility
was not a black-and-white matter, he saw two options. Firstly, the linguist could
impose a yes-or-no answer, which Hockett (1958: 323) believed to be less artificial
than it sounded, because ‘most pairs of idiolects, chosen at random, yield results
near one end of the scale or the other’. Secondly, the linguist could quantify the
matter and look for statistical solutions, a method he discussed at length, motiv-
ated by recent empirical studies in the field. In fact, in the early 1950s, American
structuralists had started to devise mutual intelligibility tests and conduct them in
actual practice, which Uriel Weinreich (1954: 397–8), too, had already appreciated
because they yielded insight into linguistic relationships, even though dividing
points resulting from such tests still remained arbitrary.
A pioneering contribution was the 1951 paper ‘Methods for determining
intelligibility among dialects of natural languages’ by the American anthropo-
logical linguist Carl F. Voegelin (1906–86) and the well-known structuralist Zellig
Harris (1909–92). Voegelin and Harris (1951: 322) recognized the possibilities of a
‘combination of eliciting and text recording’, which allowed
to take old problems in linguistic theory and interpretation and address these to
new languages, especially to American Indian languages which have the practical
advantages of being numerous, diverse, and available. One of these problems
concerns the old and difficult question of dialect versus separate language.
They identified four distinct methods used to answer this ‘old question’. Firstly,
anthropologists often simply asked informants whether two varieties were so
similar that they could be viewed as dialects, thus leaving the matter to the
subjective perception of laypeople. Secondly, linguists could quantify the degree
of similarity, ‘count sameness’ in short. The results of these two methods often
coincided, they noticed, for instance in the case of the Salish languages, spoken in
the Pacific Northwest, where independent studies following these distinct
methods resulted in similar conclusions.
DOI: 10.1093/oso/9780198845713.001.0001
  257
The second method was pioneered by Morris Swadesh (1909–67), a student of

Edward Sapir’s, who wanted to measure linguistic nearness and distance rather
than solve the language/dialect puzzle (Swadesh 1950). Swadesh did so by a
method he later dubbed ‘lexicostatistics’ (Swadesh 1952; 1955; cf. Chapter 14),
which started from a list of 100 words for each language compared. The concep-
tual pair was, however, already defined in his 1950 paper on Salish tongues, but he
distanced himself from the interpretation proposed by attributing it to ‘scholars’
in general:
Scholars commonly use dialect to refer to a local variety of a language. Language

in turn is used for a continuum of dialects such that each is mutually intelligible
with its neighbors. (Swadesh 1950: 163)
In what followed, Swadesh rejected the criterion of mutual intelligibility in

particular as too subjective.
Swadesh had focused on words, but one might as well include phonetic and
other features, Voegelin and Harris (1951: 326) argued; they did so prior to Polák,
who independently made a similar suggestion (Chapter 19, Section 19.3). This
broader perspective brought them to the third method of ‘structural status’, which
implied ascribing hierarchical values to the differences and similarities between
language varieties. Yet this structural criterion was not enough, and a fourth
method should be introduced: that of mutual intelligibility, testing whether one
informant understood another. This method was all the more necessary since
speakers of highly similar varieties could experience great difficulties in under-
standing each other because certain near-identical formal elements could receive
very different meanings and functions.
In conclusion, Voegelin and Harris were looking for objective criteria
grounded in linguistic data, on the one hand, and in intelligibility tests, on the
other. By way of check, the results of these methods could be compared to speaker
perceptions. Soon linguists from Voegelin and Harris’s network and others put
this intelligibility testing to practice in studying Amerindian tongues and other
understudied language families, for instance those in New Guinea.¹ Scholars at last
started to empirically investigate a criterion which for almost four centuries had
been loosely suggested and intuitively assessed, ever since the work of Johannes
Goropius Becanus in the second half of the sixteenth century (see Chapter 8,
Sections 8.4–5).
¹ See e.g. Hickerson et al. (1952); Pierce (1952). Cf. Casad’s (1974) handbook for mutual intelligi-
bility testing. For New Guinea see Wurm and Laycock (1961).
258  
20.1 A criterion making career
The criteria launched by anthropologists and structuralists in the early 1950s were
highly influential. Mutual intelligibility testing, in particular, was to remain a
major guiding principle among linguists who continued to use the conceptual
pair, even up to this day, with quantitative methods being increasingly adopted
(e.g. Hammarström 2008; Gooskens in Boberg et al. 2018: 206). Most notably, it is
the main criterion of Ethnologue, the widely known online language catalogue
purporting to list every recognized language—7,111 at the time of writing—in
which it follows the ISO 939-3 standard for language names, pioneered by
Ethnologue.² The use of the mutual intelligibility criterion has been defended by
claiming that the resulting classification usually agrees with the ideas of specialists
(Hammarström 2005). Rather infelicitously, the criterion is mixed by Ethnologue
with such language-external factors as ‘the existence of a common literature or of a
common ethnolinguistic identity’ and ‘the existence of well-established distinct
ethnolinguistic identities’. The result is a hybrid framework, which, although
flexible, is not as objective as it might be. Another well-known catalogue of
the world’s known languages, Glottolog, is more strict, since it takes mutual
intelligibility as its sole criterion in distinguishing languages, which leads to the
figure of 7,596 languages at the time of writing.³
Problems with the criterion of mutual intelligibility were recognized at an early
stage (cf. Hammarström 2008; Gooskens in Boberg et al. 2018: Chapter 11). It is,
for instance, possible that a speaker of one variety understands a speaker of
another better than vice versa, which linguists call asymmetry. Intelligibility can
moreover be incomplete and is, as Charles Hockett already pointed out, not a
simple yes-or-no question. Previous knowledge on the part of a speaker or listener
can moreover blur the results of intelligibility tests, as can (a lack of) good will and
other preconceptions (e.g. Rubin 1992). Acknowledgement of such and other
problems with mutual intelligibility does not necessarily imply rejecting the
validity of the conceptual pair, however, as is abundantly clear from the recent
Handbook of dialectology, in which Charlotte Gooskens states:
Reflecting on the objections to intelligibility as the most important criterion for

distinguishing between dialects and languages (the Abstand criterion), we see
that it is not unproblematic to use this criterion. More research is needed before
we will be able to establish when two varieties are so different that they are no
longer mutually intelligible, and which linguistic factors play a role. Since there is
no universally accepted criterion for distinguishing a language from a dialect, the
² See https://www.ethnologue.com/about; https://iso639-3.sil.org/about/scope#Individual%20lan

guages, both last accessed 10 January 2020.
³ See https://glottolog.org/glottolog/glottologinformation, last accessed 10 January 2020.
.    259
examples provided in the rest of this chapter are from language varieties that are
traditionally referred to as dialects as well as closely related varieties that are
mostly referred to as languages. (Gooskens in Boberg et al. 2018: 208)
Despite Gooskens’s optimism that a mutual intelligibility threshold might be

discovered, it is hard to see how this will work, especially considering the
numerous problems involving the criterion, not in the least because it is often
a question of gradation and the countless tests conducted ever since the 1950s
have not yet revealed such a threshold. What is more, the fact that mutual
intelligibility assessments can be distorted owing to language-external factors,
such as a speaker’s reluctance to understand the speech of a hostile commu-
nity, makes its validity even more questionable. Finally, on a more fundamen-
tal level, is mutual intelligibility not an effect of underlying similarities in
language itself and lack thereof not a result of differences? Would it not be
more opportune to study the phenomenon of linguistic diversity at its core, in
language itself?
20.2 Measuring linguistic distance
Such questions have led some linguists to continue rather on Morris Swadesh’s
path and measure the degree of differences between speech forms (‘count same-
ness’). They consider this method more objective as it starts from language itself
and not from a secondary effect, even though the results of both methods,
measuring linguistic distance and testing mutual intelligibility, often coincide,
studies find (e.g. Biggs 1957: 61). The secondary nature of mutual intelligibility
has been argued by the American linguist Frederick B. Agard (1907–93), who
formulated in 1971 nine postulates for distinguishing a language from a dialect. In
an account that is principally structuralist but assumes the generative split
between deep and surface structure, Agard (1971: 7) rejected the mutual intelli-
gibility criterion as ‘a matter of probability’. Instead, he attached great importance
to lexical correspondences and, in second place, phonemic inventory and gram-
matical phenomena.
Agard’s main argument was that structural criteria should be sought to distin-
guish between language and dialect, that is to say, in speech itself and not, for
instance, in its speakers. His concern with the conceptual pair was fuelled by the
loose usage of laypeople, sociologists, and sociolinguists and ‘their impressionistic
criteria’ (Agard 1971: 6). Language-external circumstances, motivated by political,
literary, or social factors, should be of no importance for the descriptive linguist.
In ‘a scientific distinction’, the focus should be on actual language and—note the
structuralist terminology—its diasystems, not on secondary circumstances, which
included mutual intelligibility:
260  
To sum up: Mutual intelligibility either across languages or within languages is of

itself unmeasurable but can nevertheless be a function of measurable phono-
logical, grammatical and lexical factors. (Agard 1971: 20)
Agard’s postulates involved a hierarchical ordering of linguistic criteria to

differentiate language from dialect. It was indeed more common among structur-
alist linguists to propose weighing the different criteria in determining the precise
status of a speech form. Usually, however, phonemic similarities or differences
were then invoked as the prime criterion (Gordon in Boberg et al. 2018: 83–4).
Agard did not propose measuring linguistic distance, as Swadesh had done, but
shared with him the desire to find stricter linguistic criteria. Perhaps Agard
considered it an impossible or undesirable task to quantify his language/dialect
framework.
In recent times, however, the availability of advanced software has made it
easier to apply statistical methods to linguistic issues in general and to the
problematic conceptual pair in particular. One of the latest and highly technical
contributions on the language/dialect opposition takes Swadesh’s approach of
measuring linguistic distance as its starting point. The paper, published in 2018,
relies on a statistical analysis of basic vocabulary items from twenty-two language
groups.⁴ Its presumed main author, Søren Wichmann, starting from the belief that
the conceptual pair is still useful and should be redefined scientifically, wants to
show that lexical, and hence linguistic, distance tends to be ‘bimodally distributed’;
this means that there are usually two peaks around which the results tend to
situate themselves, which might suggest that there is something like a language
cluster and a dialect cluster (see Figure 20.1).
Figure 20.1 Example of a typical

g/t bimodal distribution on a scale
Results in-between are rare and, hence, constitute interesting problems for
scholars. Wichmann (2016: 5) feels confident enough to advance a ‘universal
cut-off ’ point
which may be used to distinguish pairs of dialects from pairs of languages. This
distinguishing criterion is easily applied and it was arrived at by an entirely
⁴ Ran and Wichmann (2018), the published version in Mandarin. Wichmann (2016) is the English
draft version I quote here.
.  261
objective procedure which can be both replicated and revised on an evolving

dataset.
Based on his study of twenty-two language groups, the cut-off point coincides
according to Wichmann with a so-called Levenshtein distance normalized (LDN)
value of 0.49. Levenshtein distance, also known as edit distance, refers in linguis-
tics to the number of changes needed to arrive from a word in one language to the
word denoting the same concept in another. The qualification ‘normalized’
designates that the Levenshtein distance for a word pair has been ‘divided by
the length of the longer of the two words’ (Wichmann 2016: 2). Establishing the
cut-off point allows, Wichmann proceeds, the determination of whether the
language count in catalogues such as Ethnologue is correct. His conclusion is
that ‘Ethnologue tends to overdifferentiate, so the number of languages counted in
this catalogue would be too high’ (Wichmann 2016: 6).
Wichmann leaves it to the reader to conclude on the basis of his proposed cut-
off point that, for instance, Indonesian–Malaysian, Bosnian–Croatian, Hindi–Urdu,
and Standard German–Swiss German are dialect pairs, whereas Danish–Swedish,
Dongshan Chinese–Fuzhou Chinese, Cairo Arabic–Moroccan Arabic, and
Catalan–Spanish are pairs of related languages (see Wichmann 2016: 6–7, esp.
Table 2). The discussion is rounded off by an interesting suggestion which reinte-
grates mutual intelligibility into the debate on the language/dialect distinction as a
key contextual factor rather than a criterion:
what is it about the way that speakers interact which allows us to distinguish
languages and dialects? A possible explanation is that there is a threshold of
mutual intelligibility where language varieties will be influencing one another if
they are below the threshold but will cease to influence one another if they are
above it. If mutual intelligibility between variety A and B is impeded completely
speakers may take recourse to just the more prestigious of the two, if not some
third language, leaving A and B to drift apart more rapidly than would the case if
both A and B were used for communication between the two groups.
(Wichmann 2016: 7)
By also including the values of a variant of the LDN, Wichmann likewise calcu-
lates the period of time required to form new languages, which, he argues,
coincides with a window of roughly 934 to 1195 years.
20.3 Conclusion
In conclusion, there might be, after all, an empirical measure to distinguish

between languages and dialects, even though it seems that further exploration
262  
and fine-tuning of this statistical method is required, for instance, by incorporat-

ing non-lexical evidence, too, and by according relative weights to individual
features, as Carl Voegelin and Zellig Harris already suggested over sixty years
ago and Frederick Agard likewise proposed in his postulates. If in such more
comprehensive studies the data still tend to be bimodally distributed, it will be safe
to conclude that the language/dialect pair as a metalinguistic abstraction is backed
by the majority of the linguistic evidence. Still, there will always exist less numer-
ous in-between situations, which, however, should perhaps not be viewed as
invalidating the conceptual distinction but should be studied as intriguing excep-
tions in a transition zone between the language and the dialect cluster.
21
Between two extremes
Generative and sociolinguistic interpretations
The heyday of structuralism was already coming to an end as Uriel Weinreich and
his colleagues were attempting to initiate a dialectological turn in mainstream
language studies and narrow the gap with traditional dialectology. In 1957, the
paradigm-changing Syntactic Structures by one of Zellig Harris’s students, Noam
Chomsky, was published, initiating a tradition of linguistic research that usually
left little room for research on dialectal variation. Still, the problem was not
entirely neglected, and some generativists tried to frame it within their new theory.
Was there any room for the language/dialect distinction? And if so, how was it
incorporated into generative theory?
21.1 Generative grammar: no country for old dialects?
Early generative linguists conceived of language as consisting of two layers: the

underlying set of rules, for which man had an innate capacity, and the way in
which they were transformed on the surface into spoken language.¹ This division
between deep and surface structure constituted the core of generative-
transformational approaches to language at the outset. As generativists focused
on discovering the principles of universal grammar, they tried to reduce language
and its enormous variation to universally applicable ground rules. Overcoming
variation was key, and language differences, especially small-scale ones such as
those among dialects, were principally treated as minor variations.
Revealing is what Brian Newton (1972: 5) wrote at the outset of his monograph,
The generative interpretation of dialect: A study of Modern Greek phonology,
published in 1972:
Because dialects arise from a more or less uniform language it is possible to show
that they can, for the most part, be described in terms of a common set of
underlying forms; variation is introduced by the phonological processes on these
forms.
¹ On the history of generative grammar see e.g. Freidin (2012). See also Hinskens in Boberg et al.
(2018: Chapter 5) for generativism in relation to dialectology.
DOI: 10.1093/oso/9780198845713.001.0001
264   
According to Newton (1972: 3), this dialectal variation could be due either to a
difference in the order of some underlying rules in two given dialects, or in the
failure of a dialect to undergo a certain rule, which he linked to historical
developments. In doing so, Newton was following the generativist interpretation
of dialects current since Morris Halle’s (1923–2018) foundational 1962 article
‘Phonology in generative grammar’ (Halle 1962; cf. Campbell 1972).
Like Polák’s structuralist conception of the language/dialect distinction, gen-
erativist interpretations remind somewhat of what I have called the Aristotelian
criterion and might well be seen as a more sophisticated variation on this theme.
Languages differed from each other considerably in grammatical structures,
whereas related dialects showed only limited differences (cf. Löffler 1974: 4).
This idea was expressed very transparently by the generative linguist Sol Saporta
(1925–2008) in 1965:
the grammatical description of a given dialect may be converted into an adequate

description of a related dialect by the addition, deletion, or reordering of a
relatively small number of rules. Indeed, it is tempting to propose that the degree
of difference between dialects is nothing more than a function of the number and
type of such changes.
(Saporta 1965: 219; for its exemplarity see Campbell 1972: 291)
This grammar and its variations in the rules between underlying and surface
forms were, however, usually devoid of any psychological reality, Lyle Campbell
(1972) argued in his critical assessment of generative dialectology, which he
entitled in Weinreichian fashion ‘Is a generative dialectology possible?’
Apart from the view that dialects were variations on one single grammar, there
was one other major strand in early generative dialectology. This alternative
approach assumed that grammars of individual dialects should be composed
and subsequently compared, which according to Campbell (1972: 294) ‘fails in
its overemphasis of the differences at the expense of the similarities, an expense
which means that this approach fails to resolve the conflict between structure and
variation’. There was no longer a basis of comparison for the two dialects, when
two separate grammars were composed which from the point of view of deep
structure were as dissimilar as those of Chinese and Quechua; in both cases, the
grammars ‘will have different underlying forms with different rules and different
surface forms’ (Campbell 1972: 293).
The focus of generative linguists has usually not been on the question of how
one might be able to distinguish languages from dialects but on how to integrate
this type of variation into their theory and reconcile it with the transformational
framework. This integration has been an issue in subsequent updates and
.      ? 265
revisions of the theory, in most of which Chomsky has played a prominent role.²
In the 1980s, the Principles and Parameters theory was formulated, which con-
ceived of universal grammar as a system of very general abstract principles.
Different choices of largely syntactic parameters allowed a limited amount of
linguistic variation. If these choices concerned major parameters, dubbed macro-
parametric variation, this resulted in language-level differences, whereas
micro-parametric variation was responsible for differences across related dialects.
In the early 1990s, the Minimalist programme was proposed as the new major
path of generative linguistic study, in which the strictly syntactic approach to
variation was abandoned and an insistence on rule and derivation supplanted the
traditional division between deep and surface structure. Generativists started to
consider now other non-syntactic factors in integrating variation into their theory,
including cognition, physiology, and society. In particular, linguists such as Sjef
Barbiers have tried to narrow the gap between generative theory and sociolinguis-
tics, much as Weinreich tried to do in the 1950s for structuralist theory and
dialectology. In the later 1990s, Optimality Theory abandoned the rule and
derivation framework, working instead with universal constraints that work on
languages and seek to produce optimal output forms. As the importance of these
constraints is different for each language, and as they can be superseded by
rivalling constraints, there is variation. Constraint ranking accounts for differ-
ences between discrete linguistic systems—i.e. language-level variation—whereas
minor variations on this ranking are usually held responsible for dialect-level
differences.
In conclusion, generative grammar has greatly struggled to incorporate lower-
level variation into their various theoretical frameworks. The focus has been on
the ideal speaker and on the transition from underlying presentations to output
forms; language is seen as a closed system, most likely because well-defined
standard languages and not dialects have been the central research object—
especially English. It would be, however, too rash to claim that the language/
dialect distinction was not a concern for generativists at all. Still, the opposition
usually does not figure prominently in their accounts, and one has to look for
traces of it in their models of language. It seems that dialect-level variation has by
and large been presented in generative approaches as the result of minor changes
in rules, parameters, or constraints; language-level differences, in turn, are the
product of major divergences in them.
² See Hinskens in Boberg et al. (2018: Chapter 5, especially 89–91), which guides my discussion here.
Cf. also Shapiro and Schiffman (1981: 50–1).
266   
21.2 Variables over systems
The shying away from superficial forms of language variation common in struc-
turalist and generativist approaches set off a strong reaction in the 1960s.
Language is a social phenomenon, and bears indelible traces of this identity,
many linguists increasingly realized. One scholar in particular contributed greatly
to promoting the study of language in its societal context: Uriel Weinreich’s
student William Labov. After a brief career as an industrial chemist, Labov
switched to linguistics in the 1960s, graduating at Columbia University. His MA
thesis on diphthong change in the English spoken on the island of Martha’s
Vineyard of 1963 became the first landmark in what is known as sociolinguistics,
soon followed by his warehouse study, in which he charted social variation in the
pronunciation of [r] among New Yorkers (see Labov 1963; 1972).
Labov succeeded in initiating and establishing sociolinguistics as a subfield of
institutionalized language studies and bringing it closer to general linguistics.
The great appeal of his approach consists in the fact that it makes the correlation
between language and society very palpable; it focuses on language in actual
usage and tries to discover how linguistic variables are motivated by social
circumstances. Even though sociolinguistics has greatly increased our under-
standing of language variation, scholars active in this strand have generally
passed over theoretical questions about what a dialect is and what distinguishes
it from a language. Linguistic variables have taken centre stage (cf. Macaulay
2010: 64). As a consequence, sociolinguists tend to follow Hockett’s judgement
that the flexibility of terms such as language and dialect is an asset rather than a
defect. It allows linguists to speak about language in its many appearances,
ranging from distinct languages such as English, national varieties such as
American English, varieties of regions as for instance Pacific Northwest
English, or varieties of a town or city, including Pittsburgh English. The
Scottish sociolinguist Ronald K. S. Macaulay (2010: 65–6), for instance, does
not consider this flexible usage problematic and even finds it preferable that such
labels match the ideas of lay speakers:
Does it make sense to consider such diverse phenomena under the single label of
“dialects”? To my mind, yes. If the term dialect is to have any coherent inter-
pretation it must refer to a form of speech based on geographical location,
preferably one that the speakers themselves would most likely admit to
using . . . The goal of dialectology should be to tabulate those features of speech
that characterize a particular locality regardless of its size.
Not unsurprisingly, sociolinguists mostly focus on variation within a language

and are therefore less concerned with the question of how to distinguish between
language and dialect. In particular, variation within cities has been a prominent
.    267
research topic among sociolinguists, since this kind of diversity had been largely
neglected by traditional dialectologists, whose attention was monopolized by rural
dialects (see e.g. Chambers and Trudgill 1998: Chapter 4). Sociolinguists’ interest
in the correlation of linguistic variables with social factors has moreover led them
to foreground extra-linguistic factors when speaking about dialect in the sense of
‘regional form of speech’. Various sociocultural, political, and ethnic circum-
stances are invoked to identify a dialect and a dialect area. Macaulay (2010: 66),
for instance, agrees with the Austrian-American linguist Hans Kurath (1972: 76)
that ‘[w]hat identifies a locality as a dialect area will depend . . . on “a combination
of socio-cultural factors” such as political domains or settlement areas’. Yet
no strict boundaries between dialect areas can be observed, as Macaulay (2010:
66 n.2), Labov (2010: 165–6), and many other sociolinguists acknowledge.
The term dialect is frequently also used, however, to refer to varieties not bound
to a specific locality, as, for instance, the ethnically defined African American
Vernacular English (Macaulay 2010: 66–7). Still, sociolinguists consider variation
in a language from a linguistic point of view, too. Labov (1973), for instance, has
expressed the idea that dialects of one language share the same grammar or, at
least, should be encompassed within one single grammar. As such, he accepts
assumptions about the existence of a so-called panlectal or polylectal grammar
advanced by generativists in the 1960s and early 1970s. This idea has subsequently
been doubted by sociolinguists such as Peter Trudgill (1983: Chapter 1).
In summary, it seems that among many sociolinguists a rather flexible usage of
the term dialect is current, tied primarily to extra-linguistic factors. This exter-
nalist perspective is related to the main method of this branch of linguistics: the
relating of language variation to language-external attributes. As a result, the
language/dialect distinction is usually not problematized extensively by sociolin-
guists. J. K. Chambers and Peter Trudgill, for instance, treat the matter only very
briefly at the outset of their handbook on dialectology, of which the following
paragraph is the most telling:
It seems, then, that while the criterion of mutual intelligibility may have some
relevance, it is not especially useful in helping us to decide what is and is not a
language. In fact, our discussion of the Scandinavian languages and German
suggests that (unless we want to change radically our everyday assumptions
about what a language is) we have to recognise that, paradoxically enough, a
“language” is not a particularly linguistic notion at all. Linguistic features obvi-
ously come into it, but it is clear that we consider Norwegian, Swedish, Danish
and German to be single languages for reasons that are as much political,
geographical, historical, sociological and cultural as linguistic. It is of course
relevant that all three Scandinavian languages have distinct, codified, standar-
dised forms, with their own orthographies, grammar books, and literatures; that
they correspond to three separate nation states; and that their speakers consider
268   
that they speak different languages. The term “language”, then, is from a linguis-
tic point of view a relatively nontechnical term.
(Chambers and Trudgill 1998: 4–5)
For this reason, Chambers and Trudgill (1998) abandon the term language and
choose to work with the term variety, because it is more neutral and has the
advantage of being flexible and widely applicable. It can be ‘appl[ied] to any
particular kind of language which we wish, for some purpose, to consider as a
single entity’. The term dialect is retained to refer to ‘varieties which are gram-
matically (and perhaps lexically) as well as phonologically different from other
varieties’, whereas accent designates ‘the way in which a speaker pronounces, and
therefore refers to a variety which is phonetically and/or phonologically different
from other varieties’ (Chambers and Trudgill 1998: 5). Different classes of vari-
ation constitute here the vague and hesitatingly proposed criterion to distinguish a
dialect from a mere accent.
21.3 Destructive dismissal
Chambers and Trudgill (1998) thus propose to abandon the conceptual pair as a
technical distinction and to use a more neutral term instead, variety, in combin-
ation with dialect and accent. Their dismissal of the language/dialect opposition is
still gentle, but it is a dismissal nonetheless. Other linguists have more pronounced
condemnations in store, as can be easily gathered from a recent paper by
Alexander Maxwell (2018). Untangling the often baffling uses and misuses of
the Weinreich witticism in modern linguistics, Maxwell (2018: 272) indicates how
some scholars have ‘exploit[ed] the witticism’s humor to denigrate the language-
dialect dichotomy and any who invoke it’. Among those cited by Maxwell, Noam
Chomsky’s (Figure 21.1) voice is particularly powerful. Where I have had to
search for traces of the language/dialect distinction in mainstream generative
theory earlier in this chapter, it now becomes clear why.
To Chomsky, apparently convinced that it is easy to distinguish languages from
each other and from dialects, the conceptual pair is non-scientific. It therefore has
no relevance for his theory of language. In his conversations with the French
poetess and linguist Mitsou Ronat (1946–84) published in 1977, he explained that
‘in my view, the concept of language is not a linguistic concept’.³ It is worthwhile to
quote here at length his objection to the conceptual pair and, more generally, to
the externalist approach current in the then young branch of sociolinguistics:
³ Chomsky (1977: 195): ‘A mon avis, la notion de langue n’est pas une notion linguistique’. I quote
from Ronat’s French version of Chomsky’s original ‘American’, as the English rendering of the French
translation (Chomsky 1979) departs from it considerably.
.   269
Figure 21.1 Noam Chomsky in 1977

Source: Wikimedia Commons / Nationaal Archief. CC BY-SA 3.0
What is Chinese? Something yellow on a map. The languages which are called
Chinese are as diverse as the Romance languages. A Chinese from the north does
not understand a Chinese from the south, etc. It is political reasons which define
Chinese. Theoretically, nothing allows one to assert that Chinese is one language
and that the Romance languages are not another. And still, no one says that
Italian and French are one single language. No linguistic argument would
support that. I would say more: what makes French a language? I suppose that
fifty years ago neighbouring villages spoke dialects of French that were suffi-
ciently different to render intelligibility poor indeed. What, then, is a language? It
is said jokingly that a language is something that has an army and a navy. It is not
a linguistic concept, nor a linguistic definition. As far as other sociolinguistic
questions are concerned, I think that it is justified to ask oneself whether they
have been posed in a way which permits answers of some depth . . . You see,
questions about languages are always tied to those of power . . . Every one of us
speaks a certain number of [idealized linguistic] systems, mixing them in an
amusing way. Since our experience is different, our system mixtures are different.
270   
Yet I do not believe that outside the reality of these systems there exists a dialect
or language reality.⁴
Chomsky continued to repeat this harsh judgement of sociolinguistics in other

writings of his in the 1980s and 1990s (see e.g. Chomsky 1986: 15; 1992: 216; 1997:
125). In fact, he has used the Weinreich quip so very often in a disparaging
manner that several subsequent scholars have wrongly attributed the saying to
the generativist (Maxwell 2018: 267).
Misusers of the witticism not only want to show the involvement of non-
linguistic aspects in language studies in a bad light, but they also dismiss the
language/dialect distinction as having no value whatsoever. Maxwell has identified
another trend among users of the witticism apart from the dismissive one. Some
linguists recognize that their field should engage with factors outside language in
studying this phenomenon but employ the witticism to keep themselves from
doing so. The phrase serves as a disclaimer: ‘there are political factors, but these do
not concern me here’. It is striking how often linguists present the conceptual pair
as an unworkable instrument because of its political connotations but nonetheless
persist in using it, sometimes even only a few sentences apart (Maxwell 2018:
272–5, with striking examples). This unease with the conceptual pair results,
Maxwell (2018: 276–8) argues, from the fact that linguists are not trained to
conduct political analysis but do feel the need to clearly categorize linguistic
diversity.
21.4 Supplementing the conceptual pair
Numerous sociolinguists, too, have questioned the validity of the conceptual pair,
but instead of simply dismissing the problem they have come up with more
constructive solutions. If the language/dialect distinction did not work in and of
itself, supplementing it should do the trick. One of the most famous and
⁴ Chomsky (1977: 195–7): ‘Qu’est-ce que le chinois? Quelque chose de jaune sur une carte. Les
langues qu’on appelle le chinois sont aussi diverses que les langues romanes. Un chinois du nord ne
comprend pas un chinois du sud, etc. Ce sont des raisons politiques qui définissent le chinois.
Théoriquement, rien ne permet d’affirmer que le chinois est une langue et que les langues romanes
n’en sont pas une autre. Et pourtant, personne ne dit que l’italien et le français sont une même langue.
Aucun argument linguistique ne soutiendrait cela. Je dirais plus : qu’est-ce qui fait du français une
langue? Je suppose qu’il y a cinquante ans, des villages voisins parlaient des dialectes du français
suffisamment différents pour que l’intelligibilité ait été tout à fait faible. Donc, qu’est-ce qu’une langue?
On dit par plaisanterie qu’une langue, c’est ce qui a une armée et une marine. Ce n’est pas un concept
linguistique, ni une définition linguistique. Quant aux autres questions de sociolinguistique, je crois
qu’il est juste de se demander si elles ont été posées d’une manière permettant des réponses de
profondeur quelconque . . . Voyez-vous, les questions de langues sont toujours liées à celles du
pouvoir . . . Chacun de nous parle un certain nombre de ces systèmes, en les mélangeant de manière
amusante. Parce que notre expérience est différente, nos mélanges de systèmes sont différents. Mais je
ne crois pas qu’en dehors de la réalité de ces systèmes, il existe une réalité dialecte, ou langue’.
.     271
influential supplementations of the traditional pair has been proposed by Heinz

Kloss (1904–87), a German specialist in minority languages who in 1952 devised a
framework separating linguistic from external factors, into which he integrated
the conceptual pair.⁵
Kloss distinguished between language by distance (Abstandsprache) and lan-
guage by elaboration (Ausbausprache). The former referred to speech forms which
were so distant from one another in linguistic terms that they could not be
considered related dialects. Kloss (1967: 30) was aware that this linguistic distance
criterion was a tricky affair but ‘for the moment . . . assume[d] that linguists are in
a position to apply final, reliable, and uniform criteria’, since he was himself more
interested in the non-linguistic, sociological side of things. He did, however,
venture some thoughts on the matter (Kloss 1978: 63–6). Substantial differences
in phonology and grammar were required to brand two speech forms separate
languages, but the principal defining characteristic was the lexicon, according to
Kloss. Related speech forms were to be viewed as dialects if they shared basic
vocabulary and if, as a result, speakers of these varieties were able to understand
one another. Kloss realized, however, that ever since Swadesh had started com-
puting basic vocabulary likeness (see Chapter 20), linguists disagreed about the
cut-off point between language and dialect. Swadesh suggested on the basis of a
diachronic study of English vocabulary that, from a synchronic perspective, 81%
of the basic vocabulary should be the same in order to speak of related dialects,
whereas later scholars proposed a lower percentage. Measuring mutual intelligi-
bility was likewise not without problems, and Kloss cited the traditional objections
to it, while pointing out that it was a secondary effect of linguistic similarities,
principally in the lexicon but also in grammar.
A language by elaboration, an ausbau language, on the other hand, ‘refers to
languages which have deliberately been reshaped so as to become vehicles of
variegated literary expression’ (Kloss 1967: 30). Kloss’s main interest was glotto-
political, and his work thus concerned these deliberately planned ausbau lan-
guages and their relationship to spoken language within society, usually dialects or
closely related tongues. Ausbau languages are often superposed on related dialects
or closely related tongues; they ‘roof ’ them, in Kloss’s now widely accepted
terminology. As a consequence, speakers of a language roofed by very dissimilar
ausbau languages might accept dialect status for their native tongue because of the
prestige attached to the roofing ausbau language. Some speech forms such as
Corsican, in turn, lack a roofing ausbau language closely related to it—Corsica
belongs to France politically but its speech is more closely related to Italian—and
can therefore be viewed as roofless dialects (Kloss 1978: 60–3).
⁵ Kloss (1952; 1978). I quote from the updated and enlarged 1978 edition as well as from Kloss
(1967), an English résumé article.
272   
Kloss’s separation of abstand from ausbau was triggered by his observation that
linguists, sociologists, and laypeople espoused different ideas about how to dis-
tinguish between language and dialect. He found this discrepancy obvious and
believed that linguists should not necessarily agree with laypeople’s views, which
were inspired by language-external circumstances (Kloss 1978: 24, 27). Even
though he focused on the ausbau aspect, he did find it useful to include the
abstand perspective in devising a typology of sociolinguistic situations involving
ausbau languages. His findings even led him to suggest that not only mutual
intelligibility should be involved in determining the relationship between speech
forms but also ‘mutual recognizability’. Speakers may not be able to understand a
related language but may feel a close connection with it because they recognize
features of their own speech in it (Kloss 1967: 36–7 and the interesting n.8). What
is more, his conceptual innovation was also intended to facilitate language count-
ing, where, however, he seems to have intermingled the originally structuralist
concept of diasystem with language-external factors such as the possession of a
literature and usage in schools (Kloss 1978: 31–3).
Kloss remained realistic, however:
We may not, however, be mistaken about the fact that even with the help of ever
so many criteria a clear decision will not always be possible about the question of
whether we are dealing with a dialect, an autonomous language, or a variety
(style) of an autonomous language.⁶
It is almost impossible to draw the boundaries between language and dialect
according to uniform viewpoints. The researchers’ measures are divergent and
can change. What is more, non-scientific, purely historical circumstances cannot
be entirely overlooked. For instance, the boundaries between language and
dialect are traditionally drawn more broadly within the German language
space than in the Slavic area.⁷
The conceptual pair was, however, not goalless, Kloss maintained, and one should
conceptualize language and dialect not as two entities divided by one line but as
separated by a broader transition zone which Kloss (1978: 35) termed ‘border
hem’ (Grenzsaum).
⁶ Kloss (1978: 33): ‘Wir dürfen uns freilich nicht darüber täuschen, das auch mit Hilfe noch so vieler
Bestimmungsmerkmale eine klare Entscheidung darüber, ob wir es mit einer Mundart zu tun haben
oder mit einer selbständigen Sprache oder mit einer Spielart (Schreibart) einer selbständigen Sprache,
nicht immer möglich sein wird’.
⁷ Kloss (1978: 34): ‘ . . . es fast unmöglich ist, die Grenze zwischen Sprache und Dialekt nach
einheitlichen Gesichtspunkten zu ziehen. Die Maßstäbe der Forscher sind verschieden und können
sich ändern. Auch können außerwissenschaftliche, rein historische Umstände nicht ganz außer acht
gelassen werden. Z.B. wird innerhalb des deutschen Sprachraumes die Grenze zwischen Sprache und
Mundart traditionell weiter gezogen als im slawischen Bereich’.
.     273
By supplementing the language/dialect distinction with sociological concepts,

Kloss had a major impact on sociolinguistics. Yet he struggled with the conceptual
pair himself, since he intermingled linguistic distance with language development
and planning in assessing the language/dialect status of a speech form, an inco-
herence of which he was perfectly aware. He moreover acknowledged that histor-
ical circumstances could interfere with conceptualizations of language and dialect.
Kloss’s contribution was, in conclusion, situated in acknowledging the importance
of non-linguistic factors in the debate on the language/dialect pair and in intro-
ducing new concepts based on the age-old distinction to capture the sociological
side of language; he did not want to provide a final answer to whether or how
languages and dialects were different in linguistic terms, even though he evinced a
genuine interest in issues of linguistic distance, too.
21.5 By way of conclusion: a quest for alternatives
Whereas Kloss supplemented the language/dialect pair, other scholars have

recently dismissed it altogether and supplanted it with an entirely different
conception of language. William A. Kretzschmar, Jr, for instance, has made a
strong case in favour of abandoning the conceptual distinction (see especially
Kretzschmar 2009). Language and dialect are ‘observational artifacts’, according
to Kretzschmar, who ties his views to those of earlier dialectologists and Saussure;
these linguists maintained that empirical study did not allow supposition of the
existence of autonomous and clearly delimited languages and dialects (see
Chapter 18, especially Section 18.5). The conceptual pair is a construct of what
he dubs ‘the linguistics of linguistic structure’, for which he proposes ‘the linguis-
tics of speech’ as the most viable alternative:
Taken together, the basic elements of speech correspond to what has been called
a “complex system” in sciences ranging from ecology and economics to physics.
Order emerges from such systems by means of self-organization, but the order
that arises from speech is not the same as what linguists study under the rubric of
linguistic structure. (Kretzschmar 2009: 4)
To Kretzschmar, language and dialect are remnants of a persisting tradition of

language studies starting from closed systems with their own internal structure.
Reasoning often occurs circularly in this tradition, since languages and dialects are
first presupposed; then, their features are studied, which results in confirming
the languages and dialects already presupposed (cf. Kretzschmar in Boberg et al.
2018: 59–60).
Kretzschmar’s alternative has no place for the centuries-old conceptual pair.
Overall, the balance among present-day linguists seems to be shifting towards
274   
dismissing it as a useful tool to chart language diversity for various reasons.

Kretzschmar, like Hugo Schuchardt, Gaston Paris, Jules Gilliéron, and Paul
Meyer, finds fault with the closed system approach and regards the conceptual
pair as an inadequate abstraction of linguistic realities. Other linguists, like Jacques
Van Keymeulen, dismiss it as a ‘sociological notion’ (see Boberg et al. 2018: 39).
Are we moving towards a final abandonment of the conceptual pair in main-
stream linguistics? The demise of the language/dialect distinction might well be
imminent in academia, but it will definitely not part without leaving a trace.
22
A gentle goodbye?
Dialect stripped for parts
Discontented with, and embarrassed by, the undesired externalist connotations of

the conceptual pair, many modern linguists harbour suspicions about it, if they
have not chosen to abandon it altogether, and the term dialect with it. Yet many
compositions and derivations based on dialect are still in vogue. Some of these,
such as subdialect and dialectology, are centuries old (see Chapter 8, Section 8.3;
Chapter 16, Section 16.1), but most were coined only in recent history, including
adjectives such as dialectal. More interestingly, the dia- element has made a career
in terms such as diatopic, diastratic, and diaphasic, a derivation technique popu-
larized by the work of the extravagant Romanist Eugenio Coseriu (1921–2002),
who wanted to capture the different dimensions of linguistic diversity with this
terminology: geography, society, and register (see e.g. Coseriu 1998).
The element -lect, in turn, has proved even more popular to refer to speech
varieties and linguistic features of divergent kinds. Like variety, lect per se has been
used as a neutral passe-partout term for any homogeneously conceived form of
speech (e.g. Fasold 1975: 29; Kloss 1978: 23; Haugen in Alinei 1980: 40). In order
to emphasize the different dimensions of linguistic diversity, lect has received
various specifiers, such as socio- and regio- to refer to a variety of a certain social
class and a certain region, respectively. Especially sociolect has been common since
its appearance in the 1960s.¹ Regiolect was introduced to avoid the negative
connotations of dialect and to highlight the geographical dimension of variation.
A somewhat older but today lesser-known coinage in this strand is communalect,
designating the speech of a specific community (e.g. Hickerson et al. 1952: 1).
The most successful derivation, however, has been idiolect, referring to the variety
of speech of an individual; this term became popular in the late 1940s, owing to its
central position in American structuralism, where it was the basic unit of linguistic
analysis. In about 1970, scholars introduced ethnolect and technolect, stressing the
racial and professional dimensions to variation, respectively. Technolect thus served
as an alternative to jargon. Apart from ethnolects, there are also multiethnolects,
urban youth languages which have their origin in different countries (e.g. Auer in
Boberg et al. 2018: 173). A couple of years after ethnolect and technolect, genderlect
¹ Frequency estimates are based on searches in English works by means of the Google Books Ngram
Viewer https://books.google.com/ngrams, last accessed 13 January 2020.
DOI: 10.1093/oso/9780198845713.001.0001
276    ?    
was coined as a label for linguistic features related to different sexes or genders. It
will be noticed that the appearance of terms like ethnolect and genderlect nicely
dovetails with the emergence of postcolonial and gender studies, no doubt the result
of a cross-fertilization with linguistics. More recent years have seen the appearance
of natiolect, which emphasizes the link between speech variation and the dimension
of the nation-state. From most of these -lect formations, adjectives in -lectal have
been derived, mirroring the pair dialect and dialectal.
Suffice to say, -lect has been a productive lexeme in recent decades in referring
to all kinds of language varieties, very often in the closed-system sense William
Kreztschmar has criticized. In the emergent field of lectometry, lect is more and
more interpreted not as a clearly delimited linguistic entity but as a ‘collection of
linguistic features that can vary along any extra-linguistic contextual dimension in
the broadest sense possible’, covering regional, stylistic, social, and perceptual
variables.² This field has come into being as a result of the broadening of the scope
of dialectometry, popularized by Jean Séguy and Hans Goebl from the 1970s
onwards and aimed at quantifying dialect similarities and their areal mapping
(see e.g. Séguy 1973 and Goebl 1982 for foundational studies).
Lectometry is currently especially popular among sociolinguists and scholars
active in the cognitive strand of language studies, which investigates the ways in
which language interacts with the human mind. Key cognitive concepts such as
salience and prototypicality are often put into practice in lectometric studies. It
has, for instance, been argued that ‘lects are not internally homogeneous, but are
rather characterized by centrality effects’; as a result, certain features of a lect are
more salient than others for language users (Geeraerts and Kristiansen 2015: 376).
The language/dialect distinction, however, is usually not problematized within
this relatively young field of enquiry, in which the term dialect is present only at
the margins. This lack of interest is most likely due to the feature-focused
approach to language prominent in such research.
Other derivations from dialect have also proved popular, but they are too
numerous to discuss in detail here. Compositions with dialect are likewise wide-
spread and include dialect area, dialect chain, dialect continuum, dialect conver-
gence and divergence, dialect levelling, dialect loss, dialect shift, and dialect term or
word. These are frequently also used by scholars renouncing dialect as an adequate
metalinguistic term. A striking example is James N. Adams, a specialist in the
Latin language and its diversification, who stated at the outset of his 2007
monograph The regional diversification of Latin 200 – 600:
“Dialect” in popular usage implies a distinct type of speech tied to a precise

locality, whereas the reality may be far more complex, with the boundaries,
² The quote is taken from the description of a recent panel on lectometry at www.ling.arts.kuleuven.
be/qlvl/lectometrypanel/, last accessed 13 January 2020.
   ?     277
regional and linguistic, far more blurred. There is something to be said for terms
vaguer than “dialect”, such as “variations, variety, diversity”. In using such terms
throughout this book I am acknowledging that we could never from the Latin
record determine the full range of local usages of any precisely demarcated
region, even if precisely demarcated linguistic regions ever existed. The point
will also be made below . . . that there is a difference between “regionalised
standard language”, an entity that may be identified in Latin, and a “primary
dialect”, something impossible to find in Latin. The lack of evidence for the latter
is a good reason for avoiding the term “dialect”. I am, however, happy with the
expression “dialect term”. (Adams 2007: 12)
Adams, on the one hand, has general objections to the popular term dialect, while,
on the other hand, renouncing its applicability to the context of ancient Latin.
This scepticism contrasts with his retaining the composition dialect term; a
tension which remains unresolved.
Not only dialect has been a productive source of new terminology; language,
too, has been stripped for parts in order to express new conceptualizations of the
human capacity for speech and its great diversity. In one interesting case, both
terms served to convey new concepts: doculect and languoid, coined with the
precise purpose of cancelling out any discussion about the language/dialect status
of a speech form. They have been proposed as part of ‘a triumvirate of concepts’
next to the term glossonym (Cysouw 2014). This tripartition has been described
most extensively in a paper of 2013, published after years of informal discussion at
the Max Planck Institute for Evolutionary Anthropology in Leipzig between
Michael Cysouw and Jeff Good, who have benefited from the input of many
other linguists, including the typologist Martin Haspelmath. Cysouw and Good
have introduced their three terms in order to formalize the language concept, a
concern triggered by the largescale language cataloguing projects of Ethnologue
and Glottolog and the problems involved in this endeavour, not least due to the
lack of rigorous definitions of language and dialect (see Cysouw and Good
2013: 331).
Languoid is defined as a ‘language-like object’ and ‘refer[s] to an entity used to
designate any (possibly hierarchical) grouping of doculects, in principle ranging
from a set of idiolects to a high-level language family’ (Cysouw and Good 2013:
347). Doculect, in turn, was coined at Haspelmath’s suggestion to denote ‘a named
linguistic variety as attested in a specific resource’, including documents written in
a specific language, studies of such a language, or simply vague references to it in,
for instance, travel diaries (Cysouw and Good 2013: 356). A specific languoid can
receive a label, the glossonym, which linguists and laypeople alike have used to
refer to it. Cysouw and Good’s terminology, circulating since 2006, was almost
immediately popular and now features not only in many linguistic publications,
especially by typologists, but also in online language catalogues, including WALS
278    ?    
Online and Glottolog.³ Their paper of 2013 was intended to present in more
explicit terms their meta-model of the world’s linguistic diversity; it articulates
their justified belief in the necessity of more accurate terminology in order to
document the great variety of languages across the earth. This search for an
improved jargon implies for them, like for many other contemporary linguists,
the abandonment of the language/dialect pair.
22.1 Farewell to the conceptual pair?
It is, however, unlikely that the conceptual pair will disappear altogether from
linguistics in the foreseeable future. Not only are its traces indelibly left in the
terminology of the discipline, but many linguists also persist in using it, either out
of habit or by conscious choice. The very existence of the conceptual pair keeps on
fuelling debate and has in recent decades even triggered entire books on the status
of a language variety. For instance, Jean-Michel Eloy addressed at length in his
1997 monograph on Picard the question whether this Gallo-Romance speech
form is a language or a dialect (see Eloy 1997). Eloy did so in direct response to
the 1992 European Charter for Regional or Minority Languages in order to gain
the rights protection offered by it (see Chapter 23, Section 23.3), even though
France has never ratified this text (cf. also Bichurina 2016 on Franco-Provençal/
Arpitan).
After providing a selective history of modern definitions of language and
dialect, with a focus on the French tradition, Eloy presented the history of
Picard, the theoretical framework of his analysis, and his morphological descrip-
tion of Picard. The evidence he collected led him to formulate three criteria for
attributing a separate status to a minority speech form, only two of which Picard
fulfilled. Firstly, speakers should be aware of the diglossia, which held true for
Picard. Secondly, this feeling of diglossia should be backed by linguistic evidence, a
criterion also met in the case of Picard. Thirdly, speakers should be willing to
elaborate the language to standard status, a willingness lacking among Picard
speakers. Eloy thus combined criteria of various origins in determining the
language/dialect status of a form of speech, as has so often occurred in the history
of language studies. Unlike his predecessors, however, he did so in a very con-
scious manner, yet without considering any problems that might be associated
with such a hybrid approach.
Not only in case studies such as Jean-Michel Eloy’s does the conceptual pair
play a prominent role, but even more so in several recent reference works on
language and linguistic diversity, not in the least in the Handbook of dialectology
³ See https://wals.info/ and https://glottolog.org/, last accessed 13 January 2020. Cf. Cysouw and
Good (2013: 332–3).
.      ? 279
of 2018, edited by, among others, Charles Boberg, a student of William Labov’s. Its
chapters are composed by specialists from various backgrounds and as such
represent an entire gamut of different attitudes towards the conceptual pair.
Several dialectologists contributing to the volume renounce the validity of the
distinction, either because it presupposes an underlying linguistic structure as a
closed system or because of its non-linguistic connotations.⁴ Others choose to
continue its usage. This choice may be implicit, for instance in presenting concrete
linguistic examples in terms of the opposition ‘standard language’/‘dialect’ (e.g.
Hinskens in Boberg et al. 2018: 95). Explicit definitions of the term dialect are no
rarity either, yet it is striking how diverse and noncommittal they are. The
following examples may suffice to prove this:
The term “dialect” is understood today to refer to a geographically delimited

form of language. (Hickey in Boberg et al. 2018: 23)
Dialect will be defined here as a geographically determined language variety
(local or regional), which is essentially only orally transmitted and is relatively
isolated from the roofing standard variety (if any).
(Van Keymeulen in Boberg et al. 2018: 39)
The focus on geographical space should be taken to imply that dialect refers to
regional varieties, as per traditional terminology.
(Kristiansen in Boberg et al. 2018: 107)
Firstly, it is of course important to clarify what we mean by dialect. In continental
European dialectology and sociolinguistics, the term is always used in opposition
to a standard language/variety. Hence, the standard is not considered a
dialect . . . The relational definition of dialect is, by definition, linked to a standard
variety. The terminology suggests—correctly—that once the particular relation-
ship between standard and dialect is established, it will result in the two varieties
impacting on each other, usually with the standard influencing the dialect more
than the reverse. (Auer in Boberg et al. 2018: 160)
The term “dialect” will be used rather loosely to refer to accent (i.e., phonetics/
phonology only) as well as in the strict sense, that is, a language subvariety
defined solely by grammatical and lexical properties.
(Watt in Boberg et al. 2018: 226)
Do these authors want to distance themselves from the conventionalist interpret-

ations they are proposing because they realize the problematic nature of the
conceptual pair? It appears so; the result is an awkward clinging to traditional
terminology for lack of any satisfying alternatives. This practice can be downright
⁴ See, respectively, Kretzschmar and Van Keymeulen in Boberg et al. (2018).

280    ?    
inconsistent when scholars explicitly dismiss the linguistic validity of the concep-
tual pair on the very same page (e.g. Van Keymeulen in Boberg et al. 2018: 39).
My selective overview in the past four chapters has made it clear that the
conceptual pair has been variously understood and assessed in mainstream
linguistics during the last century. For reasons of space, I cannot treat in this
book the various national traditions, in which the language/dialect distinction has
been tailored to a country’s specific context. This local refashioning has often
involved the integration of a third hierarchical layer into the presentation of a
nation’s linguistic diversity, consisting either in varieties resulting from conver-
gence between dialect and standard or in varieties historically not deriving from
the standard but from the same protolanguage as the standard. For instance, in
Great Britain, scholars and laypeople alike have supposed the existence of roughly
three forms of speech: the English standard language in its received pronunciation,
the standard language in a local accent, and regional dialects (Boberg et al. 2018: 4).
Many more examples of such accommodations of the conceptual pair to
national contexts can be cited. For instance, in modern Greece, a division into
standard Modern Greek, regional accents (idiómata; ιδιώματα), and dialects
deriving from the Ancient Greek Koine such as Cretan, Cypriot, and
Cappadocian is current (see e.g. Trudgill 2003: 49), whereas in Dutch-speaking
territory a variety between the standard and the dialects has been distinguished,
known as ‘interlanguage’ (tussentaal; e.g. Absillis et al. 2012). National traditions
do not always take over the terminology associated with the conceptual pair,
however. In Italy, for instance, the term dialetto is often avoided with reference to
Italo-Romance tongues such as Sicilian and Venetian, since this label might
wrongfully suggest that they derive from the Italian standard language, while
they in fact descend from Latin very much as does Tuscan, on which the standard
is based (Auer in Boberg et al. 2018: 160).
22.2 The language/dialect distinction after 1900:

the story of a love-hate relationship
The language/dialect debate in linguistics has raged most vigorously since 1900,
especially from the 1950s, when linguists began to devote entire papers to the
question. The year 1954 was pivotal in this regard, when publications by struc-
turalists from different backgrounds appeared. The conceptual pair was, however,
usually only treated in the margin of mainstream linguistic studies, since the grand
theories up until the 1960s focused on homogeneous language systems and ideal
speakers. Perhaps structural and generative linguists felt uncomfortable treating
the issue, as most of them refrained from investigating the phenomenon of
linguistic diversity in great detail. This lack of interest might explain why the
.     281
language/dialect distinction before the 1950s was treated at some length princi-
pally in introductory manuals to the field, in which the basic concepts of linguis-
tics could not be left implicit.
Whereas structuralists and generativists tended to shy away from matters of
language variation and focused on idealized systems, the phenomenon of diversity
became re-established as an object of study in the 1960s thanks to the work of
William Labov, who initiated a true social turn. Yet extralinguistic variables and
their correlation with specific linguistic features took and still take centre stage in
this field of enquiry, and sociolinguists, too, have refrained from extensively
reflecting on the language/dialect distinction. They are very much data-oriented,
like traditional dialectologists (cf. Boberg et al. 2018: 12–13). A similar conclusion
can be drawn for recent lectometric studies. This tilt towards actual linguistic data
may be related to the long-standing problem of delimiting speech forms, first
raised at length in the 1860s, as I have argued (Chapter 18). Focusing on data has
the advantage of avoiding the seemingly unsolvable conceptual puzzle. In recent
years, however, linguists like William Kretzschmar have proposed alternative
models of language in order to overcome this problem. Scholars engaged in
charting the linguistic diversity of the world are especially drawn to the problem
of linguistic delimitation and classification; this holds a fortiori for those focusing
on non-European tongues. It is, for instance, no coincidence that statistical
measuring of linguistic distance was introduced by Morris Swadesh, a specialist
in the indigenous languages of North and Central America, precisely for purposes
of classification.
Debates about the language/dialect distinction have often been of a cursory
nature in that the conceptual pair is rarely broached as the main problem of a
discussion. Indeed, although the principal topic of several major journal papers,
the issue has been the primary subject of only few conferences, let alone mono-
graphs, and this only from the 1970s onwards, when a kind of dialect renaissance
occurred (Mattheier 1983: 135). One of the most important exceptions is the
international symposium ‘On the theory of dialect’ (‘Zur Theorie des Dialekts’),
organized by Joachim Göschel at the Philipps-Universität Marburg from 5 to 10
September 1977. Many leading sociolinguists and dialectologists participated in
this event, including the Romanist of Romanian descent Eugenio Coseriu, the
American linguist Einar Haugen (1906–94), and Heinz Kloss. The proceedings of
the conference have been published in 1980 and include not only the papers
themselves but also useful summaries and, more interestingly, a written account of
the lengthy oral discussions following many of the presentations (Göschel et al.
1980). Anyone reading this fascinating volume will recognize in the papers and
especially the discussions afterwards most of the main tendencies in the modern
history of the conceptual pair which I have presented in the past few chapters. Let
me concisely recapitulate them here.
282    ?    
I maintain that five principal tendencies can be distinguished. Firstly, many

scholars do not question the conceptual pair at all and preserve this Renaissance
relic, which can be either defined in explicit terms or simply presupposed and used
without further definition. This continued usage ranges from monograph-length
studies like Jean-Michel Eloy’s over an insistence on mutual intelligibility to a
mere passing reference to speech form x as a dialect of language y. Secondly,
scholars have tried to redefine the conceptual pair, for instance to integrate it into
the theory advocated, as some generative linguists have attempted, or by using
new computational means, as in Wichmann’s lexicostatistic analysis.⁵ Thirdly,
linguists have supplemented the conceptual pair, because they feel that criteria of a
diverging nature have been intermingled in defining it. Heinz Kloss, for instance,
has influentially distinguished between linguistic distance (Abstand) and the
sociological aspect of elaboration (Ausbau) in treating language/dialect constella-
tions, whereas André Martinet opted for a simpler scheme with language, dialect,
and patois. More complicated constellations have been proposed by, among
others, Eugenio Coseriu (1980). Fourthly, linguists have proposed entirely new
models to capture linguistic diversity, most notably Kretzschmar’s (2009) con-
ception of language as a complex system in which variation is inherent as well as
Cysouw and Good’s (2013) ‘triumvirate’ of glossonym, doculect, and languoid,
devised for language classification purposes. Fifthly and finally, numerous lin-
guists have chosen a less constructive path and simply dismiss the conceptual pair
as irrelevant for language studies, for which they often invoke the Weinreich
witticism. Language/dialect disputes are not linguistic but sociopolitical problems,
linguists in this category argue. Yet downright dismissal does not necessarily entail
avoidance of the terms language and dialect, which can be found in traditional
meanings even in the work of linguists suspicious about the conceptual pair. Even
though modern scholars often approach the language/dialect distinction with
great scepticism, they have found in the controversial word dialect a prolific
source of new terminology, to which the many derivations and compositions
testify.
Age-old interpretations of the conceptual pair live on in many modern ideas,
even though scholars do not recognize this continuity. The geographical element,
most notably, is still one of the principal parameters in definitions of the term
dialect. The criterion of mutual intelligibility has risen to prominence since the
first tests in the early 1950s, even though it has proved to be almost as problematic
as the language/dialect pair itself. The association of dialect with small-scale
variation remains popular, too, for instance in generative approaches. Yet linguis-
tic differences are no longer interpreted in a binary fashion as in the traditional
Aristotelian criterion with its opposition substance/accidents, but in the wake of
⁵ See e.g. also Tillinger (2013) for a recent redefinition tailored to the linguistic context of France.
.     283
Morris Swadesh’s lexicostatistics a rich tradition of actively measuring linguistic

distance has established itself. Intuitive analyses have given way to empirical
studies. A crucial question has been that of the cut-off point: where do speech
forms become so different that they are distinct languages rather than related
dialects? A recent study in this vein has suggested that linguistic diversity is
bimodally distributed, which means that there might be something like a dialect
pole and a language pole around which speech varieties flock together. This
bimodal distribution does, however, remain a tendency with a transition zone
rather than an absolute rule without exceptions.
The reader may have noticed that in the past few chapters I have for the first time
made repeated reference to lay usages of the term dialect in opposition to
language. Scholars such as André Martinet, Uriel Weinreich, and Heinz Kloss,
to name but a few, expressed the need to arrive at a more technical understanding
of the conceptual pair, which, they noticed, non-specialists interpreted in ambigu-
ous ways. Where did these lay conceptions of the language/dialect distinction
originate? This question is all the more pressing, since I have argued in Part II that
the pair of concepts emerged as a technical metalinguistic instrument in the work
of Renaissance Hellenists in the early sixteenth century and not at all as a popular
idea. I will briefly look into the vulgarization of the language/dialect opposition in
the final chapter of this book, since it will help the reader understand why many
modern linguists have had the anxious reflex to distance themselves from the
conceptual distinction and, especially, its sociopolitical connotations.
23
Language, dialect, and the general public—
or how not to popularize knowledge
The taxonomy of linguistic description—that is, the identification and

enumeration of languages—is greatly hampered by the ambiguities
and obscurities attaching to the terms “language” and “dialect”.
Laymen naturally assume that these terms, which are both popular
and scientific in their use, refer to actual entities that are clearly
distinguishable and therefore enumerable. A typical question asked
of the linguist is: “How many languages are there in the world?” Or:
“How many dialects are there in this country?”
(Haugen 1966: 922)
Einar Haugen’s analysis is spot-on. The terms language and especially dialect have
‘carried the full burden of both scientific and popular usage’ (Haugen 1966: 924).
As I have contended in the previous chapters, numerous modern linguists are
aware of the awkward fact that they share the language/dialect distinction with the
general public. This realization has raised suspicion against the conceptual pair,
up to the point where many linguists have barred it from scientific usage,
assuming that it had its origin outside scholarship. Yet I have argued in Part II
that the conceptual pair emerged for the first time on the large scale in specialist
philological works in the early sixteenth century. How did it evolve from a
technical tool to a piece of common knowledge?
23.1 The conceptual pair popularized
RUn away on Thursday April 12 about two of the Clock in the

Morning, from the Right Honourable the Lady Lovelaces House in
Little Queen street, John Howel Footman to the said Lady, in a new
black Liver, and brown Perriwig, about 28 years old, low of Stature,
and speaking a Welch Dialect, having stole several things, and the
next night six Silver Forks, and six Silver Spoons, with the Arms of
the Family, and other Plate and things were suspected to be stole by
the said John out of the said House: whosoever shall bring him, or give
DOI: 10.1093/oso/9780198845713.001.0001
.     285
notice of him to Mr. Lion in Holborn Grocer, or discover the said Plate
shall be well rewarded for their pains.
(Anon. 1672)
This ‘advertisement’ was published in The London Gazette edition of 22 April

1672. Such wanted notices were far from uncommon in early modern newspapers.
What is surprising, however, is that the word dialect appeared in it as part of
the physical description of the wanted criminal. Traditionally included in
seventeenth-century English dictionaries of difficult words, dialect remained a
technical term for most of the early modern era. Indeed, it seems to have occurred
only sporadically outside learned discourse before the nineteenth century.
Another intriguing exception stems from late eighteenth-century Scotland. In
1783, the redaction of The Edinburgh Weekly Magazine received a query from one
of its readers: ‘What are the causes of a difference in dialect among people who
speak the same language?’ In the 11 September edition, the interested reader could
find the following answer:
VARIETY of dialect may originally and unavoidably arise, in some instances,

from difference in the conformation of the organs of speech. The pronunciation
of several consonants is affected by this circumstance. But the following causes,
which it is unnecessary to illustrate, are sufficient to account for that diversity of
speech which is observable in every nation:—Inattention to the pronunciation of
others, and the retaining of mistaken pronunciation; the licence of poets; the
affectation of novelty in curtailing, lengthening, and compounding of words;
borrowing from foreign languages, with a view to adorn or improve; the arbitrary
invention of words by those who are unacquainted with many already in use; the
rejection of vulgar, and the re-admission of obsolete terms. The effect of these
causes becomes remarkable in proportion to the want of acquaintance between
the inhabitants of distant provinces. Though, at the building of Babel, diversity of
languages was the cause, it is now the effect of separation among mankind. The
language of North America will, in a few centuries, become unintelligible to an
inhabitant of Great Britain. (Ruddiman 1783)
In the eighteenth century, the market for knowledge opened up, and many
newly created magazines and almanacs popularized scientific information.¹ These
publications found new readership in educated bourgeois circles; their members
cultivated broad interests and inundated the magazines with the most diverging
questions, to which the editors tried to formulate sound answers. Regional
language diversity, too, drew the attention of some educated laypeople, who
¹ See e.g. Lynn (2006: especially Chapter 2), focusing on eighteenth-century France.
286 , ,    
increasingly conversed in the standard languages, both in writing and orally. It is

in this light not surprising that in his answer to the above query the editor made
disparaging comments about the phenomenon of dialect, linking it to incorrect-
ness. Yet he had a very broad conception of it, since he took it to cover also
poetical properties and, above all, various kinds of lexical peculiarities. Isolation
and lack of contact enhanced these formal differences, and on this basis he
predicted the linguistic separation of American English, which was bound to
become unintelligible to the British. This prediction might imply that the editor
replying assumed that speakers of related dialects could understand one another,
whereas speakers of distinct languages could not. Yet he did not offer an explicit
criterion by which to differentiate language from dialect.
Outside Great Britain, the conceptual pair had also made its way to non-
specialists. The historian Johann Christoph Gatterer, for instance, who developed
an embryonic lexicostatistic method in the late eighteenth century (Chapter 14),
noticed that phrases such as ‘dialects of a language’ occurred ‘even in common
speech’.² Non-specialist familiarity with dialect in opposition to language increased
during the first half of the nineteenth century in the wake of Romanticism, which
reacted against the extreme rationalism of the Enlightenment and put emotion at
the forefront. Pristine and rustic dialects were put to use by Romantic writers for
various literary purposes, which included evoking pastoral scenes and character
marking.³ Historic dialects were set off from the frigidly formalized standard
language that was gaining wide acceptance. It has been maintained that early
dialect geographers such as Wenker were still very much working in this
Romantic spirit, since they, too, focused on the age-old and allegedly pure dialects
of the countryside, which came under pressure from the steadily advancing
standard languages.
From the end of the nineteenth century, modern linguists realized that lay-
people likewise had ideas about what constituted a dialect and what a language.
Georg von der Gabelentz reported extensively on popular interpretations of the
conceptual pair in the revised edition of his linguistics handbook, posthumously
published in 1902:
A language is considered the common property of a people, a Dialekt or a

Mundart the common property of a region—this may more or less correspond
to the ordinary conception. Schriftdeutsch is written and read in the entire
fatherland, preached by the chancellors, and learned in the schools: that is why
it is language. Bavarian, Swabian, Palatine, etc., to the contrary, are dialects. It will
be difficult to understand, in accordance with this viewpoint, that Plattdeutsch is
² Gatterer (1770: 13): ‘In Sprachuntersuchungen, und sonst auch, selbst im gemeinen Reden, kom-
men immer die Ausdrücke vor: verwandte Sprachen, Dialecte einer Sprache’.
³ See e.g. Hodson (2018) for dialect and Romanticism in Britain.
.     287
not simply a dialect, too. Newspapers, public bodies, clerics, and teachers, in
many cities most citizens indeed speak High German in Low Germany as well.
Dialect is spoken by the man in the overall. Educated men, from waiters and up,
take pains to speak dialect-free. In the interior, one concedes a particular
language only to the people whose speech one does not understand at all, as,
for instance, to our Sorbs of Lausitz. This way of judging things is external,
superficial, and necessarily leads to inconsistencies. The history of our days has,
however, proved that such views can also lead to fairly vexing consequences.
Since language is the common property of the people, language community thus
justifies the right to political unification, language diversity the right to break
away—in this manner the modern nationality principle judges according to the
motto: “as far as tongue x resounds”.⁴
Gabelentz thus rejected political parameters in determining the language/

dialect status of a speech form. More fruitful, he maintained, was another
popular criterion: mutual intelligibility. In fact, to Gabelentz’s mind this feature
could be the only workable touchstone by which to make language/dialect
decisions. Yet the criterion was, as it was used by laypeople, not objective and
explicit enough. He proposed to take the speech of uneducated people as the
foundation of intelligibility analysis, admitting, however, that the criterion
remained somewhat arbitrary. The conceptual pair could only be used for con-
venience, since precise definitions were impossible. A contemporary of Gabelentz
tellingly wrote:
No scientific or adequate definition can be given. For all practical purposes this
will suffice. A language is a big dialect, and a dialect is a little language.
(Meiklejohn’s 1891 Book of the English language
as quoted in Crowley 2003: 81)
⁴ Gabelentz (1901: 55): ‘Unter einer Sprache denkt man sich das Gemeingut eines Volkes, unter
einem Dialekte oder einer Mundart das Gemeingut einer Landschaft—dies dürfte so etwa der
Allerweltauffassung entsprechen. Schriftdeutsch wird im ganzen Vaterlande geschrieben und gelesen,
von den Kanzeln gepredigt, in den Schulen gelehrt: mithin ist es Sprache. Bairisch, Schwäbisch,
Pfälzisch u. s. w. dagegen sind Dialekte. Diesem Standpunkte wird es schwer zu begreifen, dass
Plattdeutsch nicht auch blos ein Dialekt ist; Zeitungen, Behörden, Geistliche und Lehrer, in vielen
Städten die meisten Bürger reden ja auch in Niederdeutschland hochdeutsch. Dialekt spricht der Mann
im Kittel; die Gebildeten, vom Kellner aufwärts, bemühen sich ‚dialektfrei‘ zu sprechen. Im Inlande
gesteht man nur den Leuten, deren Rede man gar nicht versteht, eine besondere Sprache zu, so
z. B. unsern Lausitzer Wenden. Diese Art die Dinge zu beurtheilen ist äusserlich, oberflächlich und
muss zu Inconsequenzen führen. Die Geschichte unserer Tage hat aber bewiesen, dass solche
Ansichten auch zu recht ärgerlichen Consequenzen führen können: weil die Sprache Gemeingut
des Volkes ist, so begründet Sprachgemeinschaft das Recht zur politischen Vereinigung,
Sprachverschiedenheit das Recht zur Losreissung—so urtheilt das moderne Nationalitätsprinzip nach
dem Wahlspruche: „soweit die x’sche Zunge klingt“.’ On Gabelentz’s ideas on the conceptual pair see
also Chapter 17, Section 17.4.
288 , ,    
23.2 The Weinreich witticism in context: the nation-state,

language policy, and mass education
Gabelentz procured a surprisingly thorough and insightful analysis of popular

conceptions of the language/dialect distinction. He deplored that the opposition
had been misused to claim or deny the right to political autonomy. Nationalist
sentiment had penetrated deeply into the conceptual pair. Already in the eight-
eenth century, the phenomenon of dialect had captured attention outside the
confines of academia among educated laypeople. In this stage, however, the lay
conception of the language/dialect distinction had been principally shaped by
ideas about cultivation and education, or a lack thereof. Yet in the modern period,
the conceptual pair became knowledge available to roughly all classes of society.
The channels through which the popularization of the conceptual pair occurred
demand more systematic research, but the distinctly political dimension which
entered the scene in the nineteenth century suggests in which direction to look.
Gabelentz, in fact, had already provided part of the answer in his analysis of the
linguistic side of the ‘modern nationality principle’. It was the language ideology of
modern nation-states which was responsible for politicizing the conceptual pair.
Standard languages became an instrument for nation-building. The members of a
nation should learn to speak the uniform language imposed by the state and
diffused through mass education and media. If one aspired to attain power,
mastery of the standard language was indispensable. Language was the state-
backed means of communication, dialect a regionally limited form of speech
current among citizens lacking both power and good education. Or to put it
with the words of Weinreich’s auditor: ‘A language is a dialect with an army and
navy’. To make the nation more homogeneous, dialects were sometimes actively
suppressed with a view to their eradication. They were useless in the nation-state’s
standard language ideology.
For reasons of space and clarity, I am generalizing. It is desirable to investigate
the matter for each nation-state separately, since there are differences between
them all, depending on such factors as the education system designed and applied,
the sociopolitical and pedagogical philosophy behind it, and the tenacity with
which the standard language was imposed. It occurred very early and strongly in
France, where in the aftermath of the French Revolution Henri Grégoire
(1750–1831), better known as the Abbé Grégoire, heralded a new era in national
attitudes towards dialectal variation. In his well-known report of 1794, presented
to the French National Convention, Grégoire discussed ‘the necessity and means
for annihilating the patois and universalizing the usage of the French language’.⁵
⁵ See the report’s title: Rapport sur la nécessité et les moyens d’anéantir le patois, et d’universaliser
l’usage de la langue française. Cf. Pop (1950: ., .6–7); Burke (2004: 10). On the place of this report
in the history of French nationalism see Bell (2003: 169–97). See Certeau et al. (2002) for the
.       289
Grégoire did not think it fit for a nation like France, mother of liberty, to remain at
the Tower of Babel language-wise. The patois were remnants of feudal society,
whereas a uniform national language spoken by all would serve to perpetuate the
hard-won liberty of the French Revolution. Interestingly, Grégoire adduced
American English as the main linguistic model for French. The annihilation of
the patois became a prominent item on the political agenda as an efficient means
for freeing the nation from feudality. This radical new start involved a severe
degradation of dialects vis-à-vis the early modern period, during which the
phenomenon was not as emphatically interpreted in negative terms.
State-guided mass education, then, was most likely a factor of prime import-
ance in spreading to nearly all citizens of the state the idea that there was
something like a language/dialect distinction. This school revolution gathered
steam especially from the last decades of the nineteenth century onwards (see
Boli et al. 1985: 146, 155; Meyer et al. 1992: 145; Ferguson 2006: 23). It was
probably more decisive in the diffusion of the conceptual pair than, for instance,
mass media. Often organized by the nation-state, mass education served to spread
its ideology, a ‘package’ of which standard language ideology was a core part (May
2011: 153–4; cf. Joseph 2004: 52). Children of different social classes and back-
grounds were confronted on a large scale with the uniform linguistic norm of the
nation-state, which they were expected to learn to read, write, and eventually also
speak. They were guided in doing so by teachers who were required to draw a
sharp dividing line between the ideal of the standard language and the various
native dialects of their students (cf. Van der Horst 2008: 189). The conceptual pair,
in sum, most likely reached the masses in an educational context shaped by the
political and ideological concerns of the nation-state’s leadership. This develop-
ment deserves much closer attention, especially since it involved the political
appropriation of technical terminology coined by philologists during the
Renaissance in the first place to describe and categorize the ubiquitous phenom-
enon of linguistic diversity.
23.3 Political activations of the conceptual pair
The vulgarization of the conceptual pair in modern times had enormous conse-
quences for its interpretations, since it led to the foregrounding of language-
external and often subjective parameters. As a product of state-backed popular-
ization through mass education, the language/dialect distinction was redefined in
predominantly political and sociocultural terms. Policy dictated a sharp divide
between the uniform standard language and the diversified dialects, and the result
relationship between the French Revolution and the patois, and the many follow-up discussions of this
issue (cf. Burke 2004: 4 n.11).
290 , ,    
was a highly prescriptivist attitude towards language. Language was the precinct of
power, progression, and cultural superiority, dialect that of subordination, the
status quo, and lack of refinement. As such, dialect suffered from a severe social
stigma. In popular usage, the conceptual pair came to be defined by external
criteria alone, which makes it a very unsuccessful case of knowledge vulgarization
indeed, since most early modern scholars and modern linguists had opted either
for a mix of language-internal and external criteria or exclusively for language-
internal criteria.
The conceptual pair’s new political guise has had far-reaching consequences. At
various occasions during the last decades, the language/dialect distinction has
been the centrepiece of numerous debates, being activated in policymaking. An
already classic example is the so-called Oakland Ebonics controversy raging from
late 1996 to early 1997, with substantial follow-up discussion (see e.g. Wolfram
1998; Edwards 2009: 79–82). In a resolution of 18 December 1996, the Oakland
School Board (California, US) recognized varieties of Ebonics, or African
American Vernacular English, as ‘not merely dialects of English’ and suggested
the use of Ebonics in teaching Standard English to Afro-Americans. The result
was a fierce debate, because the resolution, using a negative variant on the ‘to differ
only in dialect’ phrase, was widely interpreted as claiming that Ebonics constituted
a separate language rather than an English dialect. It even led to Senate hearings of
linguists and non-specialists alike, including William Labov, who were asked to
present their view on the precise status of Ebonics. Tellingly, it was more popular
language ideology than linguistic study that determined the stances of laypeople.
The American linguist Walt Wolfram, for instance, described how he was often
asked whether he ‘believed’ in Ebonics, as if a religious sect were at stake rather
than a form of language (Wolfram 1998: esp. 110). Suffice to say, linguistic
diversity can be a delicate subject, especially if it amounts to language/dialect
disputes.
Perhaps more remarkable still is the European Charter for Regional or Minority
Languages, which was signed on 5 November 1992 but came into force only on
1 March 1998. The treaty was prepared by the Council of Europe (Figure 23.1), an
international organization not to be confused with the European Union, to
safeguard the languages spoken by smaller communities of its member states. In
order to clarify what was meant by ‘regional or minority languages’, the charter
offered in article 1 a definition of this phrase:
“regional or minority languages” means languages that are: i traditionally used

within a given territory of a State by nationals of that State who form a group
numerically smaller than the rest of the State’s population; and ii different from
.       291
the official language(s) of that State; it does not include either dialects of the
official language(s) of the State or the languages of migrants.⁶
The charter’s explicit exclusion of dialects forced participating member states to

make on a large-scale language/dialect decisions for speech varieties in their
territories, often with major implications. Certain minority tongues could not
enjoy privileged treatment because governments decided that they could not. For
instance, the language of the Rusyns (Ruthenians) was not recognized by Ukraine
in the framework of the charter, thus denying it regional language status solely on
the basis of the government’s political and ethnic preconceptions about the
Rusyns, who are considered to be Ukrainians speaking a Ukrainian dialect and
not a separate people with its own language. Contradictorily, Rusyn, spoken not
only in south-western Ukraine but also in surrounding countries, did obtain
regional language status in other states which have signed and ratified the charter,
including Croatia, Romania, and Slovakia. As such, the document embodies more
than anything else the problematic politicization of the conceptual pair. It there-
fore does not seem to be a pity that this charter, noble in spirit but questionable in
Figure 23.1 The Council of Europe’s headquarters in Strasbourg

Source: Wikimedia Commons. CC BY 3.0
⁶ Council of Europe (1992: 1–2). Cf. also Gorter and Cenoz (2012: 263); Van Keymeulen in Boberg
et al. (2018: 39); Gooskens in Boberg et al. (2018: 205).
292 , ,    
execution, can be denounced at any given time by the participating countries

(Council of Europe 1992: 13, Article 22).
In the last decades, the conceptual pair has played its role in non-specialist
debates also outside of politics. Discussions on the topic are often strongly
coloured ideologically and show a remarkably high degree of vagueness.
A number of such recent debates are reflected in online publications. For instance,
a blogpost of 16 February 2014, published on the website of The Economist, a
London-based weekly news magazine, explained ‘[h]ow a dialect differs from a
language’ (R. L. G. 2014). It did so in the aftermath of a debate in Hong Kong
about the precise status of Cantonese. The education department had claimed
that, contrary to English and Standard Chinese, Cantonese was not an official
language of Hong Kong. This assertion led to a debate about the status of
Cantonese: was it a separate language or merely a dialect of Chinese?
The author of the blogpost, American journalist Robert Lane Greene, tried to
offer a definitive solution to the issue. Before dismissing the popular political
criterion with reference to the Weinreich quip, he elaborated extensively on
mutual intelligibility. The latter was attributed to professional linguists and
claimed to be the most objective criterion, ‘the linguist’s common-sense defin-
ition’. The author was aware that the criterion was not absolute, but this was no
problem for its objectivity:
Linguists have a different criterion: if two related kinds of speech are so close that
speakers can have a conversation and understand each other, they are dialects of
a single language. If comprehension is difficult to impossible, they are distinct
languages. Of course, comprehensibility is not either-or, but a continuum—and
it may even be asymmetrical. Nonetheless, mutual comprehensibility is the most
objective basis for saying whether two kinds of speech are languages or dialects.
(R. L. G. 2014)
The conclusion was that Cantonese must be a distinct language, related to other
Chinese languages such as Mandarin and Shanghainese. The author added that
this ‘objective’ mutual intelligibility criterion ‘can annoy nationalists’ in China as
well as elsewhere. To further support the argument, the typical example of Danish
and Norwegian was invoked to illustrate that two mutually intelligible varieties
could be the standard languages of different nation-states. This political circum-
stance made speakers of these tongues reluctant to regard their respective stand-
ard languages as ‘only’ dialects of one and the same language. In conclusion, the
author wanted to create the impression that the interpretation of the conceptual
pair was undisputedly clear among linguists and tried to validate it as a means to
provide an allegedly objective answer to a politically thorny issue.
In recent times, the conceptual pair has been politically activated to make
various claims. In such cases, it tends to arouse strong feelings. For reasons of
.  293
convenience, I have focused on two notorious recent episodes in language policy

and an English-language blogpost, but the uses and appropriations of the con-
ceptual pair outside academic linguistics deserve a closer investigation, taking into
account various governing bodies and different languages and national contexts. It
might, moreover, be fruitful to study popular attitudes towards the language/
dialect distinction through experiments, questionnaires, or social media. The last
resource, in particular, offers many research opportunities not only to dialectolo-
gists and variationist linguists but also to scholars investigating language attitudes
and popular interpretations of widely known linguistic concepts. Especially
Twitter, the largest part of which is accessible to anyone with an Internet con-
nection, provides the researcher with many interesting and easily analysable data.
23.4 Conclusion
The conceptual pair gained some ground outside scholarly circles as early as the
seventeenth century and especially in educated lay circles during the eighteenth
century. It was, however, only with the advent of modernity that the language/
dialect distinction reached nearly all classes of society in the wake of mass
communication and especially general education instituted by the nation-state.
Language is the official means of communication of the nation-state, the refined
standard language, the language of promotion, dialect the diversified and repulsive
speech of the lower populace and the status quo. Such has been the major political
interpretation of the conceptual pair in modern times, even though in recent
decades a countermovement has done much to value dialect in more positive
terms. This popular take on the two concepts is motivated exclusively by subject-
ive, language-external parameters. As such, the language/dialect distinction has
fallen victim to the language policy and planning, the glottopolitics, of modern
nation-states, as well as to the sociolinguistic prejudices of lay speakers. It has
become a prime example of how not to popularize scholarly terminology.
The discourse of laypeople and linguists alike suggests that there has been, and
still is, a strict distinction between lay and specialist ideas, and the latter often
condemn the views of the former. Yet it is certainly not the case that ideas from
both sides are formulated in complete isolation from each other. Linguists have,
for instance, been consulted by policymakers—recall that William Labov partook
in the Oakland Ebonics Senate hearings—and have written blogposts and
recorded podcasts on the language/dialect distinction, often pointing out its
meaninglessness and political inspiration (e.g. McWhorter 2016; 2019). The
blogger for The Economist, too, confronted lay with specialist conceptions.
This dialectic interaction between laypeople, policymakers, and linguists
deserves further study, especially since it will allow linguists to trace currents of
influence from linguist to layperson as well as from layperson to linguist.
294 , ,    
Laypeople looking for an answer have, for instance, turned to linguists to find
more objective criteria, such as mutual intelligibility. Linguists opposing the pure
and correct norm to the deviant dialects, in turn, have been influenced by
subjective conceptions and the standard language ideology of the nation-state in
their prescriptivist attitudes, even though it must be granted that they are becom-
ing an increasingly rare breed. Still, Calvet (1974: 46–8) has pointed out that
sociopolitical factors underlie numerous definitions of the language/dialect dis-
tinction formulated by linguists in the twentieth century.
The strict distinction claimed to exist between lay and specialist conceptions
adumbrates the fact that there is a great diversity of ideas within each group itself.
I have illustrated this abundance of opinions at length over the past chapters for
scholarly interpretations of the language/dialect distinction. There is, however, no
uniform popular conception of it either, and this variation, too, is a matter
requiring further research. Peter Auer has recently suggested that there are two
main popular interpretations of dialect, both championing in vagueness (see
Boberg et al. 2018: 160). On the one hand, it is viewed as a pure and pristine
relic from the past, enjoying prestige because of its antiquity; this is the traditional
take on dialect cultivated also by most dialectologists up until the 1960s. Less
appreciative is, on the other hand, the conception of dialect as an inferior
deviation from the standard language. Further research can no doubt offer a
finer-grained picture of the wealth of popular ideas about the language/dialect
distinction.
I have not insisted on the ways in which the conceptual pair has served, or still
serves, in academic branches other than linguistics, which would have led me too
far. Yet the ideas of sociologists, anthropologists, historians, and the like also
deserve closer study as well as their interactions, both with lay ideas and with
views in other academic branches. That these interactions are very real, is proved
by, for instance, Louis-Jean Calvet’s Linguistics and colonialism: A short treatise on
glottophagy of 1974. In this pioneering and polemical book at the crossroads of
linguistics and postcolonial studies, Calvet analysed the ways in which Western
linguistic concepts have been carriers of dominant imperialist ideologies and have
given shape to ideas about ‘the other’ and their speech. His postcolonial perspec-
tive led him to denounce vehemently the conceptual pair not only because of the
vagueness with which it has been defined but especially because of its language-
external, political interpretations (Calvet 1974: 40–54). Imperialist discourse has
resulted in the denigration and degradation of certain speech forms to dialect
status in favour of the language of the establishment solely on the basis of power
relations and preconceptions about Western cultural superiority. To Calvet, ‘a
dialect is always nothing more than a defeated language, and the language is a
.  295
dialect which has succeeded politically’, which is why he proposed to abandon this
terminology, since the burden of its connotation was too heavy to carry.⁷
Calvet’s book, in short, suggests that cross-fertilization between different aca-
demic branches has indeed occurred; it is, for instance, not inconceivable that
postcolonialism has contributed much to exposing the political connotations of
the language/dialect distinction. Needless to say, Calvet is only one eminent
example among many, all of which deserve a more in-depth historiographic study.
⁷ Calvet (1974: 54): ‘le dialecte n’est jamais qu’une langue battue, et . . . la langue est un dialecte qui a
réussi politiquement’.
24
Language and dialect between
past and future
Terminological success, conceptual failure?
I have tried to guide the reader on the age-long journey of the distinction between
language and dialect. Its history is, however, not as long as may be suspected. The
conceptual pair does not hark back to ancient Greece, as one might guess from
the etymology of the word dialect, but rather to the sixteenth century, the period of
the Renaissance, when humanists rediscovered and appropriated Greek scholar-
ship. Yet its history, although shorter than often assumed, is still a fascinating one.
In fact, it emerged under peculiar conditions in the early sixteenth century. There
was a renewed interest in ancient Greek literature, which unlike its Latin coun-
terpart was composed in a gamut of different dialects. Also, reinforced by the
invention of the printing press and embryonic national sentiments, processes of
language standardization were gathering steam; these involved differentiating the
selected variety, the language, from other speech forms, the dialects, even though
this contrast did initially not occur in predominantly depreciatory terms for the
dialect pole.
The linguistic horizon of scholars moreover opened up. For the first time, they
had what I have dubbed a ‘multidialectal outlook’, as it was no longer exceptional
to be acquainted with the dialects of different languages, typically those of Ancient
Greek and one’s native tongue. Humanists needed concepts to bring some
structure into the enormous linguistic diversity they observed, and language was
as much as anything else the object of classification attempts in the Renaissance,
an era which witnessed a true explosion of information. The language/dialect
distinction proved a welcome classificatory tool, since it made languages as well as
dialects countable entities, an evolution embodied by nothing better than Conrad
Gessner’s 1555 language catalogue.
The conceptual pair was, in other words, a construct of its age. What Brian
Ogilvie (2006: 1) has stated about the Renaissance ‘invention of natural history’
may therefore be taken to apply to the language/dialect distinction, too: ‘it could
be the product only of a community’, in particular a community of scholars. As
opposed to Roger Bacon’s ephemeral advocating of a similar conceptual oppos-
ition in the thirteenth century, humanists were more united in their endeavour,
and their innovation met with overwhelming success. Ironically, unlike Bacon,
DOI: 10.1093/oso/9780198845713.001.0001
       297
they were unaware of the novelty of their distinction, since they assumed that they
were simply borrowing it from the greatly admired ancient Greeks, without
realizing that they were enforcing their own interpretation on the works of their
illustrious predecessors. It is telling that I have had to diagnose, so to speak, the
emergence of the language/dialect opposition by means of symptoms arising,
roughly, in the years 1500–50.
As such, the conceptual pair was a blend of ancient ideas and new insights, a
typical product of Renaissance scholarship (cf. Rice and Grafton 1994: xiii, 83).
Humanists intuitively transformed the impressionistic concept of ancient Greek
and Byzantine scholarship, diálektos as ‘particularity of tongue’, which could
acquire secondary parameters such as ‘typical of a certain tribe/region’, into a
reified and countable entity. Dialect came to be subsumed under an overarching
language concept, a process in which the initially secondary parameters gradually
gained a more prominent position. Yet because of its peculiar emergence conditions,
the conceptual pair only received explicit interpretations after several decades, from
about 1550 onwards. Still, much was left to the reader to guess, and early modern
authors could greatly vary in their conceptions of the language/dialect contrast.
Daniel Baggioni has aptly summarized this:
It would be very difficult to find a beginning of an agreement on the terms

language and dialect, which vary considerably in their meaning from one author
to the other and from one period to the other.¹
I have, however, tried to illustrate that it is nonetheless possible to observe

broader tendencies in early modern ideas. I have identified seven major interpret-
ations of the conceptual pair, which I need not repeat here. The making explicit
of how a dialect differed from a language went hand in hand with the emancipa-
tion of the distinction from Greek scholarship. From the seventeenth century
onwards, it was no longer Hellenists alone who made substantial contributions to
conceptions of the language/dialect contrast. After about 1650, once the major
early modern interpretations had been established, the focus shifted to the ways in
which related dialects differed linguistically from one another in contrast with
distinct languages, as I have argued mainly through the work of Georg Stiernhielm
and Albert Schultens. This Enlightenment rationalization of the conceptual pair
culminated in an embryonic lexicostatistic approach in the work of the
eighteenth-century Göttingen historian Johann Christoph Gatterer. At this
stage, the first critical voices about the conceptual pair and some of its interpret-
ations were also heard.
¹ Baggioni (1997: 56): ‘On serait bien en peine . . . de trouver un début d’accord sur les termes langue
et dialecte qui, d’un auteur à l’autre, d’une époque à l’autre, varient considérablement dans leur
signification’.
298       
Suffice to say, the conceptual pair had already come a long way when it was
adopted as a kind of self-evident given by scholars in the early nineteenth century,
at which time linguistics emerged as a separate scholarly discipline. Owing to the
initial focus of linguists on the interrelationships of Indo-European tongues and
their common origin, the language-historical interpretation was foregrounded.
Dialects descended from a language, like daughters from a mother. The geograph-
ical aspect rose to prominence during the 1870s in the wake of Georg Wenker’s
dialect studies and his pioneering linguistic atlas. At the same time, the results of
this new dialectological research made scholars such as Hugo Schuchardt realize
that dialects were not as countable as was commonly believed ever since the
Renaissance and that the distinction between language and dialect was above all
arbitrary. For the first time, the conceptual pair came to be questioned on a large
scale, a scepticism heard most loudly in France, especially in Paris, during the
decades around 1900.
While historical-comparative linguistics and dialectology continued to be prac-
tised throughout the twentieth century, they were superseded quite rapidly by the
grand theories of structuralism and later generative grammar, in which there was
usually little space for small-scale language variation and hence also for the
conceptual pair. It was only in the 1950s that a change was initiated, when Uriel
Weinreich dared to ask the question whether a structural dialectology was possible.
Only from this moment onwards did the debate about the language/dialect dis-
tinction start to rage in full vigour, to which the diverse journal articles entirely
devoted to the matter testify. It is telling that since the 1950s many linguists have
tried to come up with objective criteria to separate languages from dialects, very
likely reflecting an urge to counter the rise of language-external, often political and
subjective interpretations from the late nineteenth century onwards.
The booming of the debate was also tied, like in the Renaissance, to an
increased concern over counting the languages of the world, embodied in the
various online language catalogues created in recent decades. Seen from this point
of view, it is not surprising that many contributions to the debate have been
fuelled by attempts at classifying the many indigenous languages outside Europe,
where strong preconceived ideas about the language/dialect status of individual
speech forms were lacking. Even though the debate has raged most intensely in the
last seventy years, it has taken place in a rather unstructured fashion, with many
intervals and limited direct discussion.
In sum, the conceptual pair, although usually taken as an obvious and ahistorical
given, is a relic from a not-so-distant past. As yet, few historians of linguistics, let
alone linguists themselves, have acknowledged this; a rare exception is the Italian
scholar Mario Alinei (1980), who deserves full credit for his analysis of the
Renaissance as a pivotal period, even if his interpretation seems to be more socially
coloured than the evidence allows. The misleading timelessness of the language/
dialect opposition is probably enhanced by the fact that nobody seems to want to
       299
accept responsibility for it; the conceptual pair has, by consequence, become
something of an age-old orphan. Created by humanists in the sixteenth century
but attributed to the ancient Greeks, the language/dialect distinction rapidly
emancipated itself from that Greek patronage. As soon as this emancipation had
occurred, it was out there, free for everyone to use and interpret, for better or for
worse.
Symptomatically, citations of previous definitions of dialect as opposed to
language are of a remarkably low frequency, which suggests that there is a
reluctance to take part in a direct debate on the interpretation of the conceptual
pair. It is moreover telling that criteria have often been attributed to groups of
people as large as ‘laypeople’ and ‘linguists’, or vaguer still, to ‘some’ and ‘others’.
Mutual intelligibility, for instance, was hailed by Georg von der Gabelentz at the
end of the nineteenth century as a criterion of popular origin which linguists
should upgrade and make workable. In Greene’s blogpost on the website of The
Economist, exactly the opposite occurred; there, mutual intelligibility was
promoted as the number one objective criterion to distinguish dialects from
languages, which laypeople should take over from linguists (see Chapter 23).
Overall, there has been terminological rather than conceptual continuity
throughout the ages, from Greek antiquity and Byzantium through the early
modern period to modern institutionalized linguistics and lay ideas. It seems,
therefore, that in the case of dialect we have a big term rather than a big idea, since
it was not the term’s semantics but its formal outlook which has enjoyed enor-
mous success over more than two millennia. Indeed, the term dialect has
remained, but numerous interpretations have come and gone. Even if in recent
decades the term dialect has aroused suspicion among linguists, mainly owing to
the failure to develop an objective metalinguistic concept, its terminological
success continues through various derivations and compositions as well as for
the sake of tradition—or as the adage goes: old habits die hard.
Through my study of the conceptual pair’s history, I hope to have contributed
to bringing the long history of language studies back within the horizons of
modern linguists, who often assume that the history of their discipline is tanta-
mount to the history of institutionalized linguistics. This book has argued that the
deep foundations of a key conceptual distinction were laid several centuries before
the nineteenth century. Only by taking earlier developments into account can we
grasp more fully the often long and sometimes bumpy road which language
studies have taken.
The conceptual pair has had a rich and turbulent history, but does it have a future?
As elsewhere, here, too, accurate prediction is sheer impossible. Yet the fact that it
is out there, known to laypeople and linguists alike, suggests that its usage will not
decrease soon. Especially among laypeople and policymakers, the language/dialect
opposition may be expected to remain commonly employed, with all the problems
300       
attached to it, while among linguists the general suspicion of it may further
develop into an ever-stronger aversion. But it is utterly difficult to make a
confident prediction about its fate in academic circles. Even there, the conceptual
pair will probably be not abandoned altogether if only because it is so ingrained in
many linguists’ mental maps and because it has left indelible traces on linguistic
terminology. The fact remains, however, that mainstream linguists today tend to
avoid using the term dialect in contrast with language.
The main concerns are, on the one hand, the realization that language and
dialect appear to be clumsy abstractions of actual linguistic phenomena and, on
the other hand, the many undesirable sociopolitical and ideological connotations
attached to the conceptual contrast. The latter problem can, I believe, be easily
superseded if linguists are more explicit about their own usage and interpretation
by clearly defining the terms if they employ them and by abandoning bad citation
practices, which are all too common with reference to the conceptual pair (see
particularly Maxwell 2018). Explicitness and clarity are all the more indispensable,
since there are, as I have asserted, many different interpretations and connotations
attached to the terms language and dialect, which bear the load of an age-long and
complex history. Trickier is the former problem: is the gap between the reality of
language diversity and the concepts of language and dialect too wide to bridge?
Since I am first and foremost a historian of linguistics, this is not a question to
which I can hope to offer a satisfying answer. To several modern linguists, in any
case, this distance has seemed too vast, which is why they avoid the conceptual
pair almost like the plague. Only few constructive alternatives have been offered,
for instance by William Kretzschmar, who has put forward the appealing proposal
of conceptualizing language as a self-organizing complex system. This view might
indeed be closer to the truth but remains, of course, an abstraction itself. The
question then amounts to asking whether it will ever be possible to reach a
satisfying theoretical frame to grasp the inherent diversity of human language.
One of the major problems with the conceptual pair is borders in a twofold
sense. Firstly, when does a variety become so different from a related variety that it
is a different language rather than a related dialect? Secondly, is it at all possible to
distinguish between dialects, which often flow into one another, as linguists and
dialect geographers have argued in great numbers ever since the 1860s? The
delimitation problems suggest that a strictly binary distinction of languages and
dialects is impossible to maintain. Does, then, such a division have no basis at all
in linguistic evidence? If we would accept this conclusion, we might be faced with
more disturbing ponderings. For instance, are we to discard all previous work in
which the conceptual pair is put into practice in some way or another? This
conclusion is not reasonable, since in the past many linguists assuming the
language/dialect contrast have reached valid research results. What is more, recent
statistical studies focusing on lexical variation have suggested that there are indeed
two major degrees to which speech forms tend to vary, that there might be
       301
something like a dialect and a language pole, and that linguistic diversity tends to
organize itself on two major planes (Ran and Wichmann 2018).
It is my expectation that linguists in the very near future will be able to fine-tune
this statistical approach in determining degrees of linguistic kinship and involve
in it different aspects of language, not only the lexicon but also phonology,
morphology, grammar, and semantics (cf. already Voegelin and Harris 1951:
327). If such studies would confirm the existence of a dialect and a language
pole, perhaps there might be a future for the conceptual pair. In this case, linguistic
diversity would have to be reconceptualized as being in three zones flowing into
one another: a dialect and a language cluster with an intriguing transition zone
in-between. These results may be connected to an assessment of mutual intelligi-
bility and its role: around which point of linguistic distance do two speech forms
start being mutually unintelligible? Is mutual intelligibility a conservative force,
tying two related speech forms to the dialect cluster? Does gradual loss thereof, for
instance through the disappearance of mutual contact, contribute to a shift of these
two speech forms through the transition zone to the language cluster?
Such analyses of linguistic distance would, of course, remain snapshots in time
but reproducible and testable ones. Conducting them over extended periods of
time would moreover allow us to chart the nature, degree, and pace of linguistic
change. The proposed reconceptualization would also involve taking individual
speech forms not as closed systems, as the grand theorists of the twentieth century
have usually presupposed, but rather in the complex system sense William
Kretzschmar has suggested. Such a perspective would have the advantage of
diverting the attention of the seemingly unsolvable puzzle of linguistic borders
to the different types of language variation and their causes, properties, and
functions. Anyhow, it remains desirable to keep looking for ever more accurate
models of language that acknowledge its inherent variability. This search is all the
more exigent, since it has become somewhat overshadowed by a great enthusiasm
for the specific rather than the general in current linguistic research.
Whatever happens with the conceptual pair within linguistics and whether
scholars consider it an adequate or relevant abstraction or not, it remains a fact
that it is out there for everyone to use and misuse. For this reason alone, I agree
with what André Martinet (1954) and other scholars have pointed out in the past
century: linguists cannot simply ignore the language/dialect distinction even if
they no longer see a future for it in their own discipline. The fact of the matter is
that it originated within Renaissance philology and was popularized on a large
scale owing to developments in modern history, in particular the rise of the
nation-state and general education. As a result, it has become integrated into
the wide range of concepts non-specialists can rely on to make sense of their
surroundings in everyday life. This everyday usage is why linguists should assume
responsibility for the language/dialect pair and its meaning rather than shy away
from it as if it were not theirs to engage with. In an age in which science
302       
communication is ever more important, this call to responsibility should be

welcomed with open arms as an opportunity rather than disowned as a burden.
If not, there is a risk that Max Weinreich’s auditor from the Bronx will be right,
that dialect will always be language’s poor relation, a language without an army
and navy.
References
In early modern books I have retained the original place and publisher names for reasons of
findability, but I have added between square brackets the English place names where doubt
may arise.
van der Aa, A. J. (1867). ‘Hoius (Andreas)’, in Biographisch woordenboek der Nederlanden
. . . , continued by K. J. R. van Harderwijk and Dr. G. D. J. Schotel. Haarlem: J. J. van
Brederode, 8/2: 975–7.
Absillis, Kevin, Jürgen Jaspers, and Sarah Van Hoof (2012). De manke usurpator: Over
Verkavelingsvlaams. Ghent: Academia Press.
Académie française (1694). Le dictionnaire de l’Académie françoise. Two volumes. A Paris:
chez la Veuve de Jean Baptiste Coignard . . . et chez Jean Baptiste Coignard.
Adams, James Noel (2007). The Regional Diversification of Latin 200 – 600.
Cambridge: Cambridge University Press.
Adelung, Johann Christoph (1781). Deutsche Sprachlehre. Berlin: bey Christian Friedrich
Voß und Sohn.
Adelung, Johann Christoph (1782). Umständliches Lehrgebäude der deutschen Sprache zur
Erläuterung der deutschen Sprachlehre für Schulen. Leipzig: verlegts Johann Gottlob
Immanuel Breitkopf.
Agard, Frederick B. (1971). ‘Language and Dialect: Some Tentative Postulates’, Linguistics
65: 5–24.
Ahačič, Kozma (2008). ‘ “Omnis lingua confitebitur Deo”: Writing the First Slovenian
Grammar (1584)’, Histoire Épistémologie Langage 30 (2): 93–112.
Ahrens, Heinrich Ludolf (1839). De Graecae linguae dialectis. Volume 1. Gottingae
[Göttingen]: apud Vandenhoeck et Ruprecht.
Albertus, Laurentius (1573). Teutsch Grammatick, oder Sprach-Kunst . . . Augustæ
Vindelicorum [Augsburg]: excudebat Michaël Manger.
Albrecht, Jörn (2007). Europäischer Strukturalismus: Ein forschungsgeschichtlicher
Überblick. 3., erweiterte Auflage. Tübingen: Gunter Narr.
Alinei, Mario (1980). ‘Dialect: A Dialectical Approach’, in Göschel et al. (1980: 11–42).
Alinei, Mario (1984). ‘«Dialetto»: Un concetto rinascimentale fiorentino’, in Mario Alinei,
Lingua e dialetti: Struttura, storia e geografia. Studi linguistici e semiologici. Bologna: Il
Mulino, 169–99.
Alter, Stephen G. (2005). William Dwight Whitney and the Science of Language. Baltimore,
MD & London: The Johns Hopkins University Press.
Althamer, Andreas (1536). Commentaria Germaniae in P. Cornelij Taciti Equitis Rom[ani]
libellum de situ, moribus, & populis Germanorum. Norimbergæ [Nuremberg]: apud Ioh.
Petreium.
Amerot, Adrien (1520). Compendiu[m] Græcę gra[m]matices, perspicua breuitate co[m]-
plectens, quicq[ui]d est octo partium orationis. Louanii [Leuven]: apud Theodoricum
Martinum.
Amerot, Adrien (1530). De dialectis diuersis declinationum Græcanicarum tam in uerbis
quàm nominibus . . . Parisiis [Paris]: ex officina Gerardi Morrhii.
304 
Amirav, Hagit and Hans-Martin Kirn (2011). ‘Introduction’, in Hagit Amirav and Hans-
Martin Kirn (eds), Theodorus Bibliander: De ratione communi omnium linguarum et
literarum commentarius. Travaux d’Humanisme et Renaissance 475. Geneva: Droz,
–.
Anderson, Benedict (2006). Imagined Communities: Reflections on the Origin and Spread of
Nationalism. Revised edition. London & New York: Verso.
Anon. (1672). ‘Advertisement’, The London Gazette, 22 April 1672, edition 671.
Anon. (1743). ‘Daß es nützlich und möglich sey, die niedersächsische Sprache allmählig gar
abzuschaffen’, Der deutschen Gesellschaft in Leipzig Nachrichten und Anmerkungen
welche die Sprache, Beredsamkeit und Dichtkunst der Deutschen betreffen 3: 383–99.
Anon. (1765). ‘Patois’, in Denis Diderot and Jean le Rond d’Alembert (eds), Encyclopédie,
ou dictionnaire raisonné des sciences, des arts et des métiers . . . A Neufchastel
[Neufchâtel]: chez Samuel Faulche, 12: 174.
Antesignanus, Petrus (1554). ‘De dialectis appendix’, in Institutiones linguæ Græcæ,
N. Cleonardo autore, cum scholijs P. Antesignani Rapistagnensis. Lugduni [Lyon]: apud
Matthiam Bonhomme, 11–16.
Armitage, David (2012). ‘What’s the Big Idea? Intellectual History and the Longue Durée’,
History of European Ideas 38 (4): 493–507.
Ascoli, Graziadio Isaia (1876). ‘P. Meyer e il franco-provenzale’, Archivio glottologico
italiano 2: 385–95.
Asdal, Kristin and Ingunn Moser (2012). ‘Experiments in Context and Contexting’, Science,
Technology & Human Values 37 (4): 291–306.
Ashley, Kathleen M. and Véronique Plesch (2002). ‘The Cultural Processes of “Appropriation’’,
Journal of Medieval and Early Modern Studies 32 (1): 1–15.
Auer, Peter and Jürgen Erich Schmidt, eds (2010). Language and Space: An International
Handbook of Linguistic Variation. Handbooks of Linguistics and Communication
Science 30.1. Berlin & New York: De Gruyter Mouton.
Auroux, Sylvain, ed. (1992). Histoire des idées linguistiques. Volume 2. Philosophie et
langage. Liège: Mardaga.
Auroux, Sylvain et al., eds (2001). History of the Language Sciences: An International
Handbook on the Evolution of the Study of Language from the Beginnings to the
Present. Volume 2. Handbücher zur Sprach- und Kommunikationswissenschaft 18.
Berlin & New York: Walter de Gruyter.
Ax, Wolfram (1987). ‘Quadripertita ratio: Bemerkungen zur Geschichte eines aktuellen
Kategoriensystems (Adiectio – Detractio – Transmutatio – Immutatio)’, in Daniel
J. Taylor (ed.), The History of Linguistics in the Classical Period. Studies in the
History of the Language Sciences 46. Amsterdam & Philadelphia, PA: John Benjamins,
17–40.
Ayres-Bennett, Wendy and Magali Seijido (2011). Remarques et observations sur la langue
française: Histoire et évolution d’un genre. Histoire et évolution du français 1. Paris:
Éditions Classiques Garnier.
Bachmann, Andreas [praeses] and Michael Ludovici [respondens] (1625). Περὶ τῶν
διαλέκτων διάλεξις ϕιλολογικὴ sive de linguis . . . Lipsiæ [Leipzig]: excudebat Gregorius
Ritzsch.
Bacon, Roger (1859). Opera quædam hactenus inedita, ed. J. S. Brewer. Volume 1. Rerum
Britannicarum medii aevi scriptores. London: Longman, Green, Longman, and Roberts.
Bacon, Roger (1900). The ‘Opus majus’ of Roger Bacon, ed. John Henry Bridges. Volume 1.
London, Edinburgh & Oxford: Williams and Norgate.
Bacon, Roger (1902). The Greek Grammar of Roger Bacon and a Fragment of His Hebrew
Grammar, ed. Edmond Nolan and S. A. Hirsch. Cambridge: at the University Press.
 305
Baggioni, Daniel (1997). Langues et nations en Europe. Paris: Payot.

Bayly, Anselm (1756). An introduction to languages literary and philosophical . . . Volume 1.
London: printed for John and James Rivington.
Beauzée, Nicolas (1765). ‘Langue’, in Denis Diderot and Jean le Rond d’Alembert (eds.),
Encyclopédie, ou dictionnaire raisonné des sciences, des arts et des métiers . . . A Neufchastel
[Neufchâtel]: chez Samuel Faulche, 9: 249–66.
Bell, David A. (2003). The Cult of the Nation in France: Inventing Nationalism, 1680–1800.
Cambridge, MA & London: Harvard University Press.
Bembo, Pietro (1525). Prose . . . nelle quali si ragiona della volgar lingua . . . Vinegia [Venice]:
per Giovan Tacuino.
Benediktsson, Hreinn (1980). ‘Discussion’, in Torben Thrane et al. (eds), Typology and
Genetics of Language: Proceedings of the Rask-Hjelmslev Symposium, Held at the
University of Copenhagen 3rd–5th September, 1979. Travaux du Cercle linguistique de
Copenhague 20. Copenhagen: The Linguistic Circle of Copenhagen, 17–28.
Beroaldo, Filippo (1493). Commentationes . . . in Suetonium Tranquillum. Bononiae
[Bologna]: impressit Benedictus Hectoris Bononiensis.
Besnier, Pierre (1674). La reunion des langues ou l’art de les apprendre toutes par une seule.
A Paris: chez Sebastien Mabre-Cramoisy.
Bibliander, Theodore (1542). De optimo genere grammaticorum Hebraicorum commentar-
ius. Basileæ [Basel]: per Hieronymum Curionem.
Bibliander, Theodore (1548). De ratione communi omnium linguarum & literaru[m]
commentarius. Tiguri [Zürich]: apud Christoph. Frosch.
Bichurina, Natalia (2016). Trans-Border Communities in Europe and the Emergence of
‘New’ Languages: From ‘Francoprovençal patois’ to ‘Arpitan’ and ‘Arpitania’. Perpignan:
unpublished PhD dissertation (University of Perpignan via Domitia).
Bierbach, Christine (1978). Sprache als »Fait social«: Die linguistische Theorie F. de
Saussure’s und ihr Verhältnis zu den positivistischen Sozialwissenschaften. Linguistische
Arbeiten 59. Tübingen: Max Niemeyer.
Biggs, Bruce (1957). ‘Testing Intelligibility among Yuman Languages’, International Journal
of American Linguistics 23 (2): 57–62.
Blair, Ann M. (2010). Too Much to Know: Managing Scholarly Information before the
Modern Age. New Haven, CT & London: Yale University Press.
Blench, Roger (2006). ‘Archaeology and Language: Methods and Issues’, in John Bintliff
(ed.), A Companion to Archaeology. Malden, MA, Oxford & Carlton: Blackwell, 52–74.
Bloomfield, Leonard (1933). Language. New York: Henry Holt.
Blount, Thomas (1656). Glossographia . . . London: printed by Tho[mas] Newcomb, and are
to be sold by Humphrey Moseley . . . and George Sawbridge.
Boberg, Charles, John Nerbonne, and Dominic Watt, eds (2018). The Handbook of
Dialectology. Blackwell Handbooks in Linguistics. Hoboken, NJ: Wiley Blackwell.
Bodin, Jean (1566). Methodus, ad facilem historiarum cognitionem. Parisiis [Paris]: apud
Martinum Iuuenem.
Boli, John, Francisco O. Ramirez, and John W. Meyer (1985). ‘Explaining the Origins and
Expansion of Mass Education’, Comparative Education Review 29 (2): 145–70.
Bolognesi, Giancarlo C. (1953). ‘Sul Περὶ διαλέκτων di Gregorio di Corinto’, Aevum 27 (2):
97–120.
Bonfante, Giuliano (1953–4). ‘Ideas on the Kinship of the European Languages from 1200
to 1800’, Journal of World History 1: 679–99.
Bopp, Franz (1816). Über das Conjugationssystem der Sanskritsprache in Vergleichung mit
jenem der griechischen, lateinischen, persischen und germanischen Sprache. Frankfurt am
Main: in der Andreäischen Buchhandlung.
306 
Bopp, Franz (1820). ‘Analytical comparison of the Sanskrit, Greek, Latin, and Teutonic
languages, shewing the original identity of their grammatical structure’, Annals of
Oriental literature 1: 1–65.
Borghini, Vincenzio (1971). ‘Se la diversità della lingua greca è la medesima come la
italiana’, in John Robert Woodhouse (ed.), Scritti inediti o rari sulla lingua. Collezione
di opere inedite o rare 132. Bologna: commissione per i testi di lingua, 335–47.
Borst, Arno (1957–63). Der Turmbau von Babel: Geschichte der Meinungen über Ursprung
und Vielfalt der Sprachen und Völker. Four volumes. Stuttgart: Hiersemann.
Botley, Paul (2010). Learning Greek in Western Europe, 1396–1529: Grammars, Lexica,
and Classroom Texts. Transactions of the American Philosophical Society 100.2.
Philadelphia, PA: American Philosophical Society.
Bots, Hans and Françoise Waquet (1997). La république des lettres. Paris: Belin.
Bourgain, Pascale (1989). ‘Le sens de la langue et des langues chez Roger Bacon’, in
Geneviève Contamine (ed.), Traduction et traducteurs au Moyen Âge: Actes du colloque
international du CNRS organisé à Paris, Institut de recherche et d’histoire des textes, les
26–28 mai 1986. Documents, études et répertoires [54]. Paris: Centre National de la
Recherche Scientifique, 317–31.
Bovelles, Charles de (1533). Liber de differentia vulgariu[m] linguaru[m], & Gallici sermonis
varietate. Quæ voces apud Gallos sint factitiæ & arbitrariæ, vel barbariæ: quæ item ab
origine Latina manarint. De hallucinatione Gallicanorum nominum. Parisiis [Paris]: Ex
officina Roberti Stephani.
Bréal, Michel (1897). Essai de sémantique (Science des significations). Paris: Hachette.
Brekle, Herbert E., Edeltraud Dobnig-Jülch, Hans Jürgen Höller, and Helmut Weiß, eds
(1992–2005). Bio-bibliographisches Handbuch zur Sprachwissenschaft des 18. Jahrhunderts:
Die Grammatiker, Lexikographen und Sprachtheoretiker des deutschsprachigen Raums mit
Beschreibungen ihrer Werke. Eight volumes. Tübingen: Max Niemeyer.
Brevini, Franco, ed. (1999). La poesia in dialetto: Storia e testi dalle origini al Novecento.
Volume 2. I meridiani. Milan: Arnoldo Mondadori.
Brockliss, Laurence (2003). ‘Curricula’, in Hilde de Ridder-Symoens (ed.), A History of the
University in Europe. Cambridge: Cambridge University Press, 2: 563–620.
Bruce, F. F. (1992). ‘Lycaonia’, in David Noel Freedman (ed.), The Anchor Bible Dictionary.
New York: Doubleday, 4: 420–2.
Bruña Cuevas, Manuel (1998). ‘L’enseignement de l’r français aux Espagnols (XVIe–XIXe
siècles)’, in Teresa García-Sabell Tormo et al. (eds), Les chemins du texte: VI Coloquio da
APFFUE (Santiago, 19, 20 e 21 de febreiro de 1997). Cursos e Congresos da Universidade
de Santiago de Compostela 110. Santiago de Compostela: Universidade de Santiago de
Compostela & A.P.F.F.U.E., 2: 527–39.
Brunner, Richard J. (1971). Johann Andreas Schmeller: Sprachwissenschaftler und Philologe.
Innsbrucker Beiträge zur Sprachwissenschaft 4. Innsbruck: Institut für Vergleichende
Sprachwissenschaft der Universität Innsbruck.
Buchanan, George (1582). Rerum Scoticarum historia. Edimburgi [Edinburgh]: apud
Alexandrum Arbuthnetum.
Budzinski, Laure (2012). ‘Reconsidération étymologique et historique du vocable dialecto-
logie’, Romanica Cracoviensia 12 (1): 11–19.
Burke, Peter (2004). Languages and Communities in Early Modern Europe. Cambridge:
Cambridge University Press.
Burke, Peter (2013). ‘Nationalisms and Vernaculars, 1500–1800’, in John Breuilly (ed.),
The Oxford Handbook of the History of Nationalism. Oxford: Oxford University Press,
21–35.
 307
Bursill-Hall, G. L. (1971). Speculative Grammars of the Middle Ages: The Doctrine of partes
orationis of the Modistae. Approaches to Semiotics 11. The Hague & Paris: Mouton.
Bursill-Hall, G. L. (1975). ‘The Middle Ages’, in Thomas A. Sebeok (ed.), Historiography of
Linguistics. Current Trends in Linguistics 13. The Hague & Paris: Mouton, 1: 179–230.
Calepino, Ambrogio (1502). Calepinus. Rhegii Lingobardiæ [Reggio Emilia]: i[n]dustria
presbyteri Dionysii Berthochi imp[re]ssoris.
Calvet, Louis-Jean (1974). Linguistique et colonialisme: Petit traité de glottophagie.
Bibliothèque scientifique. Paris: Payot.
Camden, William (1595). Institutio Græcæ grammatices compendiaria . . . Londini [London]:
excudebat Edm. Bollifant pro Simone Waterson.
Camerarius, Joachim (1541). ‘Prooemium in Historias Herodoti . . . ’, in Ἡροδότου λόγοι
ἐννέα, οἵπερ ἐπικαλοῦνται μοῦσαι . . . Basileæ [Basel]: in officina Heruagiana, α.2–β.4.
Campbell, Lyle (1972). ‘Is a Generative Dialectology Possible?’, Orbis: Bulletin international
de documentation linguistique 21 (2): 289–98.
Campbell, Lyle (2006). ‘Why Sir William Jones Got It All Wrong, or Jones’ Role in How to
Establish Language Families’, Anuario del Seminario de Filología Vasca Julio de Urquijo:
International Journal of Basque Linguistics and Philology 40 (1/2): 245–64.
Canini, Angelo (1555). Ἑλληνισμός . . . Parisiis [Paris]: apud Guil. Morelium.
Casad, Eugene H. (1974). Dialect Intelligibility Testing. Summer Institute of Linguistic
Publications in Linguistics and Related Fields 38. Norman, OK: Summer Institute of
Linguistics of the University of Oklahoma.
Casaubon, Méric (1650). De quatuor linguis commentationis, pars prior: quæ de lingua
Hebraica: et, de lingua Saxonica. Londini [London]: typis J. Flesher, sumptibus Ric. Mynne.
Caxton, William (1490). ‘[Preface]’, in Vergil, The boke yf Eneydos. [Westminster]:
[William Caxton], .i–ii.
Celenza, Christopher S. (2009). ‘End Game: Humanist Latin in the Late Fifteenth Century’,
in Yanick Maes et al. (eds), Latinitas Perennis. Leiden & Boston, MA: Brill, 2: 201–44.
Ceporin, Jakob (1522). Compendium Græcæ grammaticæ. Basileæ [Basel]: apud Valentinu[m]
Curionem.
Certeau, Michel de, Dominique Julia, and Jacques Revel (2002). Une politique de la langue.
La Révolution française et les patois: L’enquête de Grégoire. Nouvelle édition augmentée.
Folio: Histoire 117. Paris: Gallimard.
Chambers, J. K. and Peter Trudgill (1998). Dialectology. Second edition. Cambridge
Textbooks in Linguistics. Cambridge: Cambridge University Press.
Charlet, Jean-Louis (2013). ‘La culture grecque de Niccolò Perotti’, Cahiers de Recherches
Médiévales et Humanistes 25: 259–80.
Cheng, Eileen Ka-May (2008). The Plain and Noble Garb of Truth: Nationalism & Impartiality
in American Historical Writing, 1784–1860. Athens, GA & London: University of Georgia
Press.
Chomsky, Noam (1977). Dialogues avec Mitsou Ronat. Traduit de l’américain et présenté
par Mitsou Ronat. Dialogues. Paris: Flammarion.
Chomsky, Noam (1979). Language and Responsibility, transl. John Viertel. Based on
conversations with Mitsou Ronat. New York: Pantheon.
Chomsky, Noam (1986). Knowledge of Language: Its Nature, Origin, and Use. Convergence.
Westport, CT & London: Praeger.
Chomsky, Noam (1992). ‘Explaining Language Use’, Philosophical Topics 20 (1): 205–31.
Chomsky, Noam (1997). ‘Language from an Internalist Perspective’, in David Martel
Johnson and Christina E. Erneling (eds), The Future of the Cognitive Revolution.
New York & Oxford: Oxford University Press, 118–35.
308 
Chrysoloras, Manuel (1512). ‘Ἐρωτήματα τοῦ Χρυσολωρᾶ’, in Ἐρωτήματα τοῦ Χρυσολωρᾶ·

Περὶ ἀνωμάλων ῥημάτων . . . Venetijs [Venice]: in ædib[us] Aldi, 3–115.
Chytraeus, Nathan (1582). Nomenclator Latinosaxonicus. Rostochii [Rostock]: typis
Stephani Myliandri.
Clackson, James (2015). ‘Latinitas, Ἑλληνισμός and Standard Languages’, Studi e Saggi
Linguistici 53 (2): 309–30.
Clement of Alexandria (1550). Κλήμεντος Ἀλεξανδρέως τὰ εὑρισκόμενα ἅπαντα, ed. Pietro
Vettori. Florentiæ [Florence]: cudebat . . . Laurentius Torrentinus.
Clement of Alexandria (1551). Clementis Alexandrini omnia quae quidem extant opera . . . ,
transl. Gentien Hervet. Florentiæ [Florence]: Laur. Torrentinus . . . excudebat.
Codro, Antonio Urceo (1502). In hoc Codri uolumine haec continentur: Orationes. seu
sermones ut ipse appellabat. Epistolæ. Siluæ. Satyræ. Eglogæ. Epigrammata. Bononiæ
[Bologna]: eme[n]date ac/ curateq[ue] impressum . . . per Ioannem Antoniu[m]
Platonidem.
Colombat, Bernard (2008). ‘L’accès aux langues pérégrines dans le Mithridate de Conrad
Gessner (1555)’, Histoire Épistémologie Langage 30 (2): 71–92.
Colombat, Bernard and Manfred Peters (2009). ‘Introduction’, in Conrad Gessner,
Mithridate. Mithridates (1555), ed. Bernard Colombat and Manfred Peters. Travaux
d’Humanisme et Renaissance 452. Geneva: Droz, 11–90.
Colvin, Stephen (1999). Dialect in Aristophanes and the Politics of Language in Ancient
Greek Literature. Oxford: Clarendon Press.
Colvin, Stephen (2010). ‘Greek Dialects in the Archaic and Classical Ages’, in Egbert
J. Bakker (ed.), A Companion to the Ancient Greek Language. Blackwell Companions
to the Ancient World. Malden, MA, Oxford & Chichester: Wiley-Blackwell, 200–12.
Consani, Carlo (1991). ΔΙΑΛΕΚΤΟΣ: Contributo alla storia del concetto di ‘dialetto’. Testi
linguistici 18. Pisa: Giardini.
Consani, Carlo (2000). ‘La nozione di «lingua comune»/«varietà dialettale» nei grammatici
tardo-antichi’, in Cristina Vallini (ed.), Le parole per le parole: I logonimi nelle lingue e nel
metalinguaggio. Atti del Convegno Napoli, Istituto Universitario Orientale 18–20 dicem-
bre 1997. Rome: Il Calamo, 605–18.
Considine, John (2008a). Dictionaries in Early Modern Europe: Lexicography and the
Making of Heritage. Cambridge: Cambridge University Press.
Considine, John (2008b). ‘Did Andreas Jäger or Georg Caspar Kirchmaier Write the
Dissertation De lingua vetustissima Europae (1686)?’, Historiographia Linguistica 35
(1–2): 13–22.
Considine, John (2010). ‘Why Was Claude de Saumaise Interested in the Scythian
Hypothesis?’, Language & History 53 (2): 81–96.
Considine, John (2014). Academy Dictionaries 1600–1800. Cambridge: Cambridge
University Press.
Considine, John (2017). Small Dictionaries and Curiosity: Lexicography and Fieldwork in
Post-Medieval Europe. Oxford: Oxford University Press.
Considine, John and Toon Van Hal (2010). ‘Introduction: Classifying and Comparing
Languages in Post-Renaissance Europe (1600–1800)’, Language & History 53 (2): 63–9.
Constantin, Emile and Ferdinand de Saussure (2005). ‘Linguistique Générale (Cours de M’
Le Professeur de Saussure) Semestre d’hiver 1910–1911’, Cahiers Ferdinand de Saussure
58: 82–290.
Cooper, Thomas (1565). Thesaurus linguæ Romanæ & Britannicæ . . . Londini [London]:
excusum . . . in ædibus quondam Bertheleti . . . per Henricum VVykes.
 309
[Coquebert de Montbret], [Charles-Étienne or Eugène] (1831). ‘Essai d’un travail sur la

géographie de la langue française’, in Mélanges sur les langues, dialectes et patois;
renfermant, entre autres, une collection de versions de la parabole de l’enfant prodigue
en cent idiomes ou patois différens, presque tous de France; précédés d’un essai d’un travail
sur la géographie de la langue française. Paris: au Bureau de l’Almanach du Commerce,
5–29.
Cortelazzo, Manlio (1980). I dialetti e la dialettologia in Italia (fino al 1800). Ars linguistica:
Commentationes analyticae et criticae 4. Tübingen: Gunter Narr.
Coseriu, Eugenio (1975). ‘Stiernhielm, die rumänische Sprache und das merkwürdige
Schicksal eines Vaterunsers’, Románica 8: 7–24.
Coseriu, Eugenio (1980). ‘“Historische Sprache” und “Dialekt”’, in Göschel (1980: 106–22).
Coseriu, Eugenio (1998). ‘Le double problème des unités “dia-s” ’, Les Cahiers δια: Études
sur la diachronie et la variation linguistique 1: 9–16.
Coseriu, Eugenio and Reinhard Meisterfeld (2003). Geschichte der romanischen
Sprachwissenschaft. Volume 1. Tübingen: Gunter Narr.
Cottière, Matthieu (1646). De Hellenistis, & lingua Hellenistica, exercitationes secundariæ.
[Strasbourg]: typis Rihelianis.
Council of Europe (1992). ‘European Charter for Regional or Minority Languages’.
European Treaty Series 148. Strasbourg: Council of Europe.
Cram, David (1999). ‘Edward Lhuyd and the Doctrine of the Permutation of Letters’, in
Gerda Haßler & Peter Schmitter (eds), Sprachdiskussion und Beschreibung von
Sprachen im 17. und 18. Jahrhundert. Studium Sprachwissenschaft 32. Münster: Nodus,
317–35.
Crastone, Giovanni (1478). Λεξικὸν κατὰ στοιχείων. [Milan]: [Bono Accursio].
Crowley, Tony (2003). Standard English and the Politics of Language. Second edition.
Basingstoke & New York: Palgrave Macmillan.
Crusius, Martin (1584). Turcogræciæ libri octo. Basileæ [Basel]: per Leonardum Ostenium,
Sebastiani Henricpetri impensa.
Cysouw, Michael (2014). ‘Languoid, Doculect and Glossonym: Formalizing the Notion
“Language”’, Billet. Diversity Linguistics Comment (blog). Online. Available at https://
dlc.hypotheses.org/623. Last accessed 25 July 2020.
Cysouw, Michael and Jeff Good (2013). ‘Languoid, Doculect, and Glossonym: Formalizing
the Notion “Language”’, Language Documentation & Conservation 7 (December).
Available at http://hdl.handle.net/10125/4606. Last accessed 25 July 2020.
Dahan, Gilbert, Irène Rosier, and Luisa Valente (1995). ‘L’arabe, le grec, l’hébreu et les
vernaculaires’, in Sten Ebbesen (ed.), Sprachtheorien in Spätantike und Mittelalter.
Geschichte der Sprachtheorie 3. Tübingen: Gunter Narr, 265–321.
Debus, Karl Heinz (1998). Der Gatterer-Apparat. Patrimonia 119. Speyer: Landesarchiv.
Delbrück, Berthold (1880). Einleitung in das Sprachstudium: Ein Beitrag zur Geschichte
und Methodik der vergleichenden Sprachforschung. Bibliothek indogermanischer
Grammatiken 4. Leipzig: Breitkopf und Härtel.
Demonet, Marie-Luce (1992). Les Voix du signe: Nature et origine du langage à la
Renaissance (1480–1580). Bibliothèque littéraire de la Renaissance: Série 3, 29. Paris &
Geneva: Champion & Slatkine.
Denecker, Tim (2017). Ideas on Language in Early Latin Christianity: From Tertullian to
Isidore of Seville. Supplements to Vigiliae Christianae 142. Leiden & Boston, MA: Brill.
Desmet, Piet (1996). La linguistique naturaliste en France (1867–1922): Nature, origine et
évolution du langage. Orbis/Supplementa 6. Leuven & Paris: Peeters.
310 
Desmet, Piet, Peter Lauwers, and Pierre Swiggers (2002). ‘Le développement de la dialecto-
logie française avant et après Gilliéron’, in Peter Lauwers, Marie-Rose Simoni-
Aurembou, and Pierre Swiggers (eds), Géographie linguistique et biologie du langage:
Autour de Jules Gilliéron. Orbis/Supplementa 20. Leuven & Paris: Peeters, 17–64.
De Laet, Johannes (1643). Notæ ad dissertationem Hugonis Grotii de origine gentium
Americanarum: et observationes aliquot ad meliorem indaginem difficillimæ illius
quæstionis. Amstelodami [Amsterdam]: apud Ludovicum Elzivirium.
Dickey, Eleanor (2007). Ancient Greek Scholarship: A Guide to Finding, Reading, and
Understanding Scholia, Commentaries, Lexica, and Grammatical Treatises, from Their
Beginnings to the Byzantine Period. American Philological Association Classical
Resources Series 7. Oxford & New York: Oxford University Press.
Dictionnaire de Trévoux (1721). Dictionnaire universel François et Latin . . . Imprimé à
Trevoux et se vend à Paris: chez Florentin Delaulne . . .
Dionisotti, Carlo (1968). Gli umanisti e il volgare fra Quattro e Cinquecento. Bibliotechina
del Saggiatore 29. Florence: Felice Le Monnier.
Dohna, Lothar Graf zu (1964). ‘Gatterer, Johann Christoph’, in Neue Deutsche Biographie.
Berlin: Duncker & Humblot, 6: 89–91.
Droixhe, Daniel (1978). La linguistique et l’appel de l’histoire (1600–1800): Rationalisme et
révolutions positivistes. Langue et cultures 10. Geneva: Droz.
Drusius, Johannes (1616). Ad uoces Ebraicas Noui Testamenti commentarius duplex . . . , ed.
Abel Curiander. Franekeræ Frisiorum [Franeker]: excudebat Fredericus Heynsius.
Edwards, John (2009). Language and Identity: An Introduction. Key Topics in
Sociolinguistics. Cambridge: Cambridge University Press.
Eichhorn, Johann Gottfried (1779). ‘Ueber die verschiedenen Mundarten der Arabischen
Sprache’, in John Richardson, Abhandlung über Sprachen, Litteratur und Gebräuche
morgenländischer Völker, transl. Friedrich Federau. Leipzig: in der Weygandschen
Buchhandlung, 3–72.
Eisenstein, Elizabeth L. (1966). ‘Clio and Chronos: An Essay on the Making and Breaking of
History-Book Time’, History and Theory 6: 36–64.
Elffers, Els (2008). ‘Georg von der Gabelentz and the Rise of General Linguistics’, in Lo van
Driel and Theo Janssen (eds), Ontheven aan de tijd: Linguïstisch-historische studies voor
Jan Noordegraaf bij zijn zestigste verjaardag. Amsterdam & Münster: Stichting
Neerlandistiek & Nodus, 191–200.
Eliot, Thomas (1538). The dictionary of syr Thomas Eliot knyght. Londini [London]: in
ædibus Thomæ Bertheleti typis impress[um].
Eloy, Jean-Michel (1997). La constitution du picard: Une approche de la notion de langue.
Bibliothèque des Cahiers de l’Institut de Linguistique de Louvain 90. Louvain-la-Neuve:
Peeters.
Eloy, Jean-Michel (1998). ‘Sur l’inventaire des variétés d’oïl de France: Un processus de
naissance de langues?’, Travaux de linguistique 9: 3–17.
Engler, Rudolf (2000). ‘La géographie linguistique’, in Sylvain Auroux (ed.), Histoire des
idées linguistiques. Bruxelles: Mardaga, 3: 239–52.
Erasmus, Desiderius (1519). In Nouum Testamentum . . . denuo recognitum, annotationes,
ingenti nuper accessione per autorem locupletatæ. Basileæ [Basel]: apud Ioannem Frobenium.
Erasmus, Desiderius (1528). De recta Latini Græciq́[ue] sermonis pronu[n]tiatione . . .
dialogus . . . Apud inclytam Basilæam [Basel]: in officina Frobeniana.
Eriksson, Manne (1971). ‘Sven Hof ’, in Svenskt biografiskt lexikon. Online. Available at
https://sok.riksarkivet.se/Sbl/Presentation.aspx?id=13681. Last accessed 10 August 2020.
Eskhult, Josef (2015). ‘Albert Schultens (1686–1750) and Primeval Language: The Crisis of a
Tradition and the Turning Point of a Discourse’, in Gerda Haßler (ed.), Metasprachliche
 311
Reflexion und Diskontinuität: Wendepunkte—Krisenzeiten—Umbrüche. Münster: Nodus,

72–94.
Eskhult, Josef (2018). ‘Vulgar Latin as an Emergent Concept in the Italian Renaissance
(1435–1601): Its Ancient and Medieval Prehistory and Its Emergence and Development
in Renaissance Linguistic Thought’, Journal of Latin Linguistics 17 (2): 191–230.
Eskhult, Josef (2020). ‘The Concept of Dialect in the Works of Three Seventeenth-Century
Linguistic Theorists, Georg Stiernhielm, Claudius Salmasius and Christian Ravius: An
Exploratory Survey of Their Intellectual Contexts, Defining Criteria, and Classificatory
Aims’, in Van Rooy (2020b).
Eskhult, Josef (forthcoming a). Albert Schultens as Teacher and Innovator of Comparative
Hebrew Language Studies: Exploratory Survey of Scholarly Authorship and Textual
Edition of Unpublished Manuscripts with Introduction, Translation and Reception
Studies.
Eskhult, Josef (forthcoming b). Georg Stiernhielm (1598–1672) on the Structure, Function
and History of Language: Textual Edition of His Theoretical Linguistic Writings with
Introduction and Reception Studies. Studia Latina Upsaliensia. Uppsala: Uppsala
Universitet.
Estienne, Henri (1579). Proiect du livre intitulé De la precellence du langage françois.
A Paris: par Mamert Patisson.
Estienne, Henri (1582). Hypomneses de Gall[ica] lingua, peregrinis eam discentibus
necessariæ . . . [Geneva]: [Henri Estienne].
Estienne, Henri (1587). Dialogus, de bene instituendis Græcæ linguæ studiis. [Geneva]:
[Henri Estienne].
Estienne, Robert (1531). Dictionarium, seu Latinæ linguæ thesaurus . . . Parisiis [Paris]: ex
officina Roberti Stephani.
Facius, Johann Friedrich (1782). Compendium dialectorum Graecarum . . . Norimbergae
[Nuremberg]: sumtibus E. C. Grattenaueri.
Fasold, Ralph W. (1975). ‘The Bailey Wave Model: A Dynamic Quantitative Paradigm’, in
Ralph W. Fasold and Roger W. Shuy (eds), Analyzing Variation in Language: Papers
from the Second Colloquium on New Ways of Analyzing Variation. Washington, DC:
Georgetown University Press, 27–58.
Fausel, Andrea, Jochen Hafner, and Franziska Küenzlen (2002). ‘Sprechen über
Sprache und Geschichte: Mittelalterliche und humanistische Entwürfe zur
Sprachgeschichtsschreibung’, in Cora Dietl and Dörte Helschinger (eds), Ars und
Scientia im Mittelalter und in der Frühen Neuzeit: Ergebnisse interdisziplinärer
Forschung. Georg Wieland zum 65. Geburtstag. Tübingen & Basel: A. Francke, 193–214.
Ferguson, Gibson (2006). Language Planning and Education. Edinburgh Textbooks in
Applied Linguistics. Edinburgh: Edinburgh University Press.
Ferté, H. (1902). Rollin: Sa vie, ses œuvres et l’université de son temps. Paris: Hachette.
Fischer, Hans (1966). Conrad Gessner (26. März 1516–13. Dezember 1565): Leben und
Werk. Neujahrsblatt 168. Zürich: Naturforschenden Gesellschaft in Zürich.
Formigari, Lia (2004). A History of Language Philosophies. Studies in the History of the
Language Sciences 105. Amsterdam & Philadelphia, PA: John Benjamins.
Franck, Jakob (1878). ‘Fulda, Friedrich Karl’, in Allgemeine Deutsche Biographie. Munich &
Leipzig: Duncker & Humblot, 8: 192.
Frederickx, Eddy and Toon Van Hal (2015). Johannes Goropius Becanus (1519–1573):
Brabants arts en taalfanaat. Hilversum: Verloren.
Freidin, Robert (2012). ‘A Brief History of Generative Grammar’, in Gillian Russell and
Delia Graff Fara (eds), The Routledge Companion to Philosophy of Language. New York
& London: Routledge, 895–916.
312 
Fulda, Friedrich Carl (1773). Ueber die beiden Hauptdialecte der Teutschen Sprachen: Eine
Preisschrift . . . Leipzig: bey Bernhard Christoph Breitkopf und Sohn.
Furetière, Antoine (1690). Dictionaire universel, contenant generalement tous les mots
françois . . . Three volumes. A la Haye, et à Rotterdam [The Hague & Rotterdam]: chez
Arnout & Reinier Leers.
Gabelentz, Georg von der (1901). Die Sprachwissenschaft, ihre Aufgaben, Methoden und
bisherigen Ergebnisse, ed. Albrecht Graf von der Schulenburg. Zweite, vermehrte und
verbesserte Auflage. Leipzig: Chr. Herm. Tauchnitz.
Galiani, Ferdinando (1779). Del dialetto napoletano. Napoli [Naples]: per Vincenzo
Mazzola-Vocola.
Gatterer, Johann Christoph (1770). ‘Vorrede, von der historischen Benutzung der
Sprachen’, in Fortsetzung der Algemeinen Welthistorie durch eine Geselschaft von
Gelehrten in Teutschland und Engeland ausgefertiget. Halle: bey Johann Justinus
Gebauer, 34: 3–16.
Gatterer, Johann Christoph (1771). Einleitung in die synchronistische Universalhistorie zur
Erläuterung seiner synchronistischen Tabellen. Göttingen: im Verlag der Wittwe
Vandenhoek.
Gaza, Theodore (1495). In hoc uolumine haec insunt. Theodori Introductiuæ gra[m]matices
libri quatuor . . . Venetiis [Venice]: impressum . . . in ædibus Aldi Romani.
Gedike, Friedrich (1779). Gedanken über Purismus und Sprachbereicherung. Berlin: gedrukt
bei George Jakob Dekker.
Gedike, Friedrich (1782). ‘Ueber Dialekte, besonders die griechischen’, Berlinsches Magazin
der Wissenschaften und Künste 1 (2): 1–26.
Gedike, Friedrich (1794). ‘Über deutsche Dialekte: Erste Vorlesung’, Beiträge zur deutschen
Sprachkunde: Vorgelesen in der Königlichen Akademie der Wissenschaften zu Berlin
1: 292–332.
Geeraerts, Dirk and Gitte Kristiansen (2015). ‘Variationist Linguistics’, in Ewa Dabrowska
and Dagmar Divjak (eds), Handbook of Cognitive Linguistics. Handbücher zur Sprach-
und Kommunikationswissenschaft 39. Berlin & Boston, MA: De Gruyter Mouton, 366–89.
Geijer, Erik Gustaf (1850). ‘Äreminne öfver Georg Stjernhjelm . . . : Uppläst i Svenska
Akademien d. 20 Dec. 1840’, in Svenska akademiens handlingar ifrån år 1796.
Stockholm: P. A. Norstedt & Söner, 23: 151–71.
Gerald of Wales (1868). Itinerarium Kambriæ, et descriptio Kambriæ, ed. James F. Dimock.
Giraldi Cambrensis Opera 6. London: Longmans, Green, Reader, and Dyer.
Gessner, Conrad (1543). ‘De utilitate ac præstantia Græcæ linguæ, in omni genere stu-
diorum, ad candidos lectores præfatio’, in Lexicon Græcolatinum. Basileæ [Basel]: per
Hieronymum Curionem, .2–7.
Gessner, Conrad (1551). Historiæ animalium lib. I de quadrupedibus uiuiparis. Tiguri
[Zürich]: apud Christ. Froschouerum.
Gessner, Conrad (1555). Mithridates: De differentiis linguarum tum ueterum tum quæ
hodie apud diuersas nationes in toto orbe terraru[m] in usu sunt . . . obseruationes.
Tiguri [Zürich]: excudebat Froschouerus.
Gessner, Conrad (1561). ‘Ad lectorem præfatio super hoc dictionario Germanicolatino . . . ’,
in Josua Maaler, Die Teütsch spraach . . . Dictionarium Germanicolatinum nouum. Tiguri
[Zürich]: excudebat Christophorus Froschouerus, *.3–7.
Giard, Luce (1984). ‘Du latin médiéval au pluriel des langues, le tournant de la Renaissance’,
Histoire Épistémologie Langage 6 (1): 35–55.
Giard, Luce (1992). ‘L’entrée en lice des vernaculaires’, in Auroux (1992: 206–25).
 313
Gierl, Martin (2012). Geschichte als präzisierte Wissenschaft: Johann Christoph Gatterer und
die Historiographie des 18. Jahrhunderts im ganzen Umfang. Fundamenta Historica 4.
Stuttgart & Bad Cannstatt: frommann-holzboog.
Gierl, Martin (2013). ‘Change of Paradigm as a Squabble between Institutions: The Institute
of Historical Sciences, the Society of Sciences, and the Separation of Cultural and Natural
Sciences in Göttingen in the Second Half of the Eighteenth Century’, in André
Holenstein et al. (eds), Scholars in Action: The Practice of Knowledge and the Figure of
the Savant in the 18th Century. Leiden & Boston, MA: Brill, 1: 267–87.
Giese, Albert (1837). Ueber den Aeolischen Dialekt: Zwei Bücher. Berlin: in der
G. Finckeschen Buchhandlung.
Gill, Alexander (1619). Logonomia Anglica. Qua gentis sermo facilius addiscitur. Londini
[London]: excudit Iohannes Beale.
Gill, Alexander (1972). Logonomia Anglica (1619), ed. Bror Danielsson and Arvid
Gabrielson, transl. Robin C. Alston. Two volumes. Acta Universitatis Stockholmiensis:
Stockholm Studies in English 26–7. Stockholm: Almqvist & Wiksell.
Gipper, Sonja (2014). ‘From Inferential to Mirative: An Interaction-Based Account of an
Emerging Semantic Extension’, in Evie Coussé and Ferdinand von Mengden (eds),
Usage-Based Approaches to Language Change. Studies in Functional and Structural
Linguistics 69. Amsterdam & Philadelphia, PA: John Benjamins, 83–116.
Gitner, Adam (2018). ‘Sardismos: A Rhetorical Term for Bilingual or Plurilingual
Interaction?’, The Classical Quarterly 68 (2): 689–704.
Glarean, Heinrich (1538). In C. Iulii Caesaris clariss. Roman. imperatoris Commentarios de
bello Gallico, ac ciuili . . . annotationes. Lugduni [Lyon]: apud Seb. Gryphium.
Goebl, Hans (1982). Dialektometrie: Prinzipien und Methoden des Einsatzes der
Numerischen Taxonomie im Bereich der Dialektgeographie. Österreichische Akademie
der Wissenschaften, Philosophisch-historische Klasse, Denkschriften 157. Vienna:
Verlag der Österreichischen Akademie der Wissenschaften.
Goebl, Hans (1983). ‘ “Stammbaum” und “Welle”: Vergleichende Betrachtungen aus
numerisch-taxonomischer Sicht’, Zeitschrift für Sprachwissenschaft 2 (1): 3–44.
Goebl, Hans (2003). ‘Graziadio Isaia Ascoli, Carlo Battisti e il ladino: Breve controstoria di
una pietra dello scandalo della linguistica a cavallo tra Otto- e Novecento’, in Antonio
Trampus and Ulrike Kindl (eds), I linguaggi e la storia. Bologna: Il Mulino, 273–98.
Goldhagen, Hermann (1752). ‘Appendix I. De dialectis’, in Institutiones linguae Latinae et
Graecae . . . Mannhemii [Mannheim]: ex typographejo electorali aulico apud Nicolaum
Pierron, 32–4 [second sequence of pagination].
Goropius Becanus, Johannes (1580). Opera . . . hactenus in lucem non edita . . . Antuerpiæ
[Antwerp]: excudebat Christophorus Plantinus.
Gorter, Durk and Jasone Cenoz (2012). ‘Legal Rights of Linguistic Minorities in the
European Union’, in Peter M. Tiersma and Lawrence M. Solan (eds), The Oxford
Handbook of Language and Law. Oxford: Oxford University Press, 261–71.
Göschel, Joachim, Pavle Ivić, and Kurt Kehr, eds (1980). Dialekt und Dialektologie:
Ergebnisse des internationalen Symposions ‘Zur Theorie des Dialekts’ Marburg/Lahn,
5.–10. September 1977. Zeitschrift für Dialektologie und Linguistik: Beihefte. Neue
Folge 26. Wiesbaden: Franz Steiner.
Göschel, Joachim, Norbert Nail, and Gaston Van der Elst, eds (1976). Zur Theorie des
Dialekts: Aufsätze aus 100 Jahren Forschung mit biographischen Anmerkungen zu den
Autoren. Zeitschrift für Dialektologie und Linguistik: Beihefte. Neue Folge 16.
Wiesbaden: Franz Steiner.
314 
Grafton, Anthony (1983). Joseph Scaliger: A Study in the History of Classical Scholarship.
Volume 1. Oxford-Warburg Studies. Oxford: Clarendon Press.
Grafton, Anthony (2007). What Was History? The Art of History in Early Modern Europe.
Cambridge: Cambridge University Press.
Grafton, Anthony (2009). ‘Codex in Crisis: The Book Dematerializes’, in Anthony Grafton,
Worlds Made by Words: Scholarship and Community in the Modern West. Cambridge,
MA & London: Harvard University Press, 288–324.
Gretser, Jakob (1593). Institutionum de octo partibus orationis, syntaxi et prosodia
Græcorum, libri tres. Ingolstadii [Ingolstadt]: excudebat Dauid Sartorius.
Grimm, Jacob (1819). Deutsche Grammatik. Volume 1. Göttingen: in der Dieterichschen
Buchhandlung.
Grinberg, Martine (2006). Écrire les coutumes. Paris: Presses Universitaires de France.
Groddeck, Benjamin [praeses] and Michael Gottlieb Treuge [respondens] (1747).
Commentatio academica de natura dialectorum ad linguam Hebraicam et Arabicam
applicata . . . Vitembergae [Wittenberg]: typis Schlomachianis.
Guillén de Brocar, Arnao (1514). ‘Introductio qua[m] breuissima ad Grecas litteras’, in
Arnao Guillén de Brocar (ed.), Nouum Testamentum Grece & Latine in Academia
Complutensi nouiter impressum. [Alcalá de Henares]: [Arnao Guillén de Brocar],
a.i–g.iii.
Guldi, Jo and David Armitage (2015). The History Manifesto. Cambridge: Cambridge
University Press.
Haas, Johann Gottfried (1780). Grichische Grammatik kurz und deutlich insonderheit den
Anfängern zum Besten abgefasset. Leipzig: im Schwickertschen Verlage.
Haas, Walter (1994). Provinzialwörter: Deutsche Idiotismensammlungen des 18. Jahrhunderts.
Historische Wortforschung: Untersuchungen zur Sprach- und Kulturgeschichte des
Deutschen in seinen europäischen Bezügen 3. Berlin & New York: Walter de Gruyter.
Hackett, Jeremiah (2013). ‘Roger Bacon’, in Edward N. Zalta (ed.), The Stanford
Encyclopedia of Philosophy. Online. Available at https://plato.stanford.edu/archives/
spr2015/entries/roger-bacon/. Last accessed 14 January 2020.
Hainsworth, J. B. (1967). ‘Greek Views of Greek Dialectology’, Transactions of the
Philological Society 65: 62–76.
Halle, Morris (1962). ‘Phonology in Generative Grammar’, Word 18 (1–3): 54–72.
Haloinus, Georgius (1978 [1533]). De restauratione linguae Latinae libri III, ed.
Constantinus Matheeussen. Bibliotheca scriptorum Graecorum et Romanorum
Teubneriana. Leipzig: B. G. Teubner.
Hammarström, Harald (2005). ‘LINGUIST List 16.2637: General Linguistics: Gordon
(2005)’, The LINGUIST List. 12 September 2005. Online. Available at https://
linguistlist.org/issues/16/16-2637.html. Last accessed 14 January 2020.
Hammarström, Harald (2008). ‘Counting Languages in Dialect Continua Using the
Criterion of Mutual Intelligibility’, Journal of Quantitative Linguistics 15 (1): 34–45.
Harris, Roy and Talbot J. Taylor (1997). Landmarks in Linguistic Thought I: The Western
Tradition from Socrates to Saussure. Second edition. Routledge History of Linguistic
Thought Series. London & New York: Routledge.
Haßler, Gerda (2009a). ‘Analogie’, in Haßler and Neis (2009: .658–74).
Haßler, Gerda (2009b). ‘Normierung’, in Haßler and Neis (2009: .674–718).
Haßler, Gerda and Cordula Neis, eds (2009). Lexikon sprachtheoretischer Grundbegriffe des
17. und 18. Jahrhunderts. Two volumes. Berlin & New York: Walter de Gruyter.
Haugen, Einar (1966). ‘Dialect, Language, Nation’, American Anthropologist 68 (4):
922–35.
Helladius, Alexander (1714). Status praesens ecclesiae Graecae . . . [Altdorf]: [s. n.].
 315
Hickerson, Harold, Glen D. Turner, and Nancy P. Hickerson (1952). ‘Testing Procedures
for Estimating Transfer of Information among Iroquois Dialects and Languages’,
International Journal of American Linguistics 18 (1): 1–8.
Hickey, Raymond (2010). ‘Attitudes and Concerns in Eighteenth-Century English’, in
Raymond Hickey (ed.), Eighteenth-Century English: Ideology and Change. Studies in
English Language. Cambridge: Cambridge University Press, 1–20.
Hinskens, Frans, Peter Auer, and Paul Kerswill (2005). ‘The Study of Dialect Convergence
and Divergence: Conceptual and Methodological Considerations’, in Peter Auer et al.
(eds), Dialect Change: Convergence and Divergence in European Languages. Cambridge:
Cambridge University Press, 1–48.
Hirschi, Caspar (2012). The Origins of Nationalism: An Alternative History from Ancient
Rome to Early Modern Germany. Cambridge: Cambridge University Press.
Hockett, Charles Francis (1958). A Course in Modern Linguistics. New York: Macmillan.
Hodson, Jane (2018). ‘Literary Uses of Dialect’, in David Duff (ed.), The Oxford Handbook
of British Romanticism. Oxford: Oxford University Press, 513–28.
Hof, Sven (1772). Dialectus Vestrogothica, ad illustrationem aliquam linguae Suecanae,
veteris et hodiernae, dissertatione philologica et vocabulorum Vestrogothicorum indice
explanata. Stockholmiae [Stockholm]: typis & impensis Joh. Arv. Carlbohm.
Hofmann, Konrad (1883). ‘Roger Bacon über die französischen und englischen
Mundarten’, Romanische Forschungen 1 (3): 427–8.
Hoius, Andreas (1620). ‘Dissertatiuncula, de linguæ Græcæ dialectorum sedibus et colo-
niis’, in Andreas Hoius, Orthoepeia, siue de Germana ac recta linguæ Græcæ, et obiter
Latinæ, pronuntiatione. Duaci [Douai]: ex officina Ioannis Bogardi, 95–104.
van der Horst, Joop (2008). Het einde van de standaardtaal: Een wisseling van Europese
taalcultuur. Amsterdam: J. M. Meulenhoff.
Hottinger, Johann Heinrich (1661). Etymologicum Orientale; sive Lexicon harmonicum
ἑπτάγλωττον . . . Francofurti [Frankfurt am Main]: sumptibus Joh. Wilhelmi Ammonij,
& Wilhelmi Serlini.
Hovdhaugen, Even (1990). ‘Una et Eadem: Some Observations on Roger Bacon’s Greek
Grammar’, in G. L. Bursill-Hall et al. (eds), De ortu grammaticae: Studies in Medieval
Grammar and Linguistic Theory in Memory of Jan Pinborg. Studies in the History of the
Language Sciences 43. Amsterdam & Philadelphia, PA: John Benjamins, 117–31.
Hovdhaugen, Even, ed. (1996). And the Word Was God: Missionary Linguistics and
Missionary Grammar. Münster: Nodus.
Hovdhaugen, Even, Fred Karlsson, Carol Henriksen, and Bengt Sigurd (2000). The History
of Linguistics in the Nordic Countries. Helsinki: Societas Scientiarum Fennica.
Hoven, René (1985). Bibliographie de trois auteurs de grammaires grecques contemporains
de Nicolas Clénard: Adrien Amerot, Arnold Oridryus, Jean Varennius. Livre – Idées –
Société 7. Aubel: P. M. Gason.
Howell, James (1642). Instructions for forreine travel . . . London: printed by T. B. for
Humphrey Mosley.
Howell, James (1650). Epistolæ Ho-Elianæ: Familiar letters domestic and forren . . . The
second edition . . . London: printed by W. H. for Humphrey Moseley.
[Howes], [Thomas] (1776). Critical observations on books, antient and modern. Volume 1.
London: printed by W. Richardson.
Howlet, Richard (1552). Abcedarium Anglicolatinum. Londini [London]: ex officina
Gulielmi Riddel.
Hoyt, David L. (2006). ‘Dialects of Modernization in France and Italy, 1865–1900’, in David
L. Hoyt and Karen Oslund (eds), The Study of Language and the Politics of Community in
Global Context. Lanham et al.: Lexington Books, 85–118.
316 
Hugo Schuchardt Archiv. Online. Available at http://schuchardt.uni-graz.at/. Last accessed

14 January 2020.
Hüllen, Werner (2001). ‘Reflections on Language in the Renaissance’, in Martin
Haspelmath et al. (eds), Language Typology and Language Universals . . . Handbooks of
Linguistics and Communication Science 20. Berlin & New York: Walter de Gruyter, 1:
210–22.
Hyde, Thomas (1767 [1691]). ‘Oratio de linguæ Arabicæ antiquitate, præstantia et utilitate’,
in Gregory Sharpe (ed.), Syntagma dissertationum quas olim auctor doctissimus Thomas
Hyde . . . separatim edidit . . . Oxonii [Oxford]: e Typographeo Clarendoniano, 2: 451–9.
Iggers, Georg G. (2011). ‘Introduction’, in Leopold von Ranke, The Theory and Practice of
History, ed. Georg G. Iggers, transl. Wilma A. Iggers. London & New York: Routledge,
xi–xlv.
Impey, Oliver and Arthur MacGregor, eds (2001). The Origins of Museums: The Cabinet of
Curiosities in Sixteenth and Seventeenth Century Europe. London: House of Stratus.
Jablonski, Paul Ernst (1714). Disquisitio de lingua Lycaonica . . . Berolini [Berlin]: literis
Wesselianis.
Jankowsky, Kurt R. (1972). The Neogrammarians: A Re-Evaluation of Their Place in the
Development of Linguistic Science. Janua Linguarum: Series Minor 116. The Hague &
Paris: Mouton.
Jenkins, Gary W. (2006). John Jewel and the English National Church: The Dilemmas of an
Erastian Reformer. Aldershot & Burlington: Ashgate.
Jones, William (1799). The works of Sir William Jones. London: printed for G. G. and
J. Robinson . . . and R. H. Evans . . . .
Jones, William Jervis (2001). ‘Early Dialectology, Etymology and Language History in
German-Speaking Countries’, in Auroux et al. (2001: 1105–15).
de Jonge, Henk J. (1980). De bestudering van het Nieuwe Testament aan de
Noordnederlandse universiteiten en het Remonstrants Seminarie van 1575 tot 1700.
Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen, Afdeling
Letterkunde, Nieuwe Reeks 106. Amsterdam, Oxford & New York: Noord-Hollandse
Uitgevers Maatschappij.
Joseph, John E. (1987). Eloquence and Power: The Rise of Language Standards and Standard
Languages. Open Linguistics Series. London: Frances Pinter.
Joseph, John E. (1996). ‘ “Undoubtedly a Powerful Influence”: Victor Henry’s Antinomies
Linguistiques (1896), with an Annotated Translation of the First Chapter’, Language &
Communication 16 (2): 117–44.
Joseph, John E. (2004). Language and Identity: National, Ethnic, Religious. Basingstoke &
New York: Palgrave Macmillan.
Joseph, John E. (2006). Language and Politics. Edinburgh Textbooks in Applied Linguistics.
Edinburgh: Edinburgh University Press.
Joseph, John E. (2012). Saussure. Oxford & New York: Oxford University Press.
Junius, Franciscus (1665). Quatuor D. N. Jesu Christi euangeliorum versiones perantiquæ
duæ, Gothica scil[icet] et Anglo-Saxonica . . . Dordrechti [Dordrecht]: typis & sumptibus
Junianis. Excudebant Henricus & Joannes Essæi.
Kals, Jan Willem (1752). Dissertatio philologica de linguæ Hebrææ natalibus punctisque
vocalibus ad eam docendam discendamque summe necessariis. Volume 1. Oxonii
[Oxford]: e Theatro Sheldoniano, impensis Joannis Barrett, et prostant apud
J. Baldwin . . . & apud Tho. Merrill.
Kamusella, Tomasz (2015). Creating Languages in Central Europe during the Last
Millennium. Basingstoke & New York: Palgrave Macmillan.
 317
Kamusella, Tomasz (2016). ‘The History of the Normative Opposition of “Language versus
Dialect”: From Its Graeco-Latin Origin to Central Europe’s Ethnolinguistic Nation-
States’, Colloquia Humanistica 5: 164–88.
ten Kate, Lambert (1723). Aenleiding tot de kennisse van het verhevene deel der
Nederduitsche sprake. Two volumes. Tot Amsterdam: by Rudolph en Gerard Wetstein.
Keil, Heinrich, ed. (1855–80). Grammatici Latini. Seven volumes. Lipsiae [Leipzig]: in
aedibus B. G. Teubneri.
Kibbee, Douglas and Marcus Keller (2018). ‘Introduction’, in Scipion Dupleix, Liberté de la
langue françoise dans sa pureté, ed. Douglas Kibbee and Marcus Keller. Remarques et
observations sur la langue française 2. Paris: Classiques Garnier, 7–167.
Kirchmaier, Georg Caspar [praeses] and Johannes Crusius [respondens] (1684). De origini-
bus & caussis literaturæ Græcæ, ejusdemq́[ue] dialectis . . . Wittenbergæ [Wittenberg]:
typis Christiani Schrödteri.
Kirchmaier, Georg Caspar [praeses] and Andreas Jäger [respondens] (1686). De lingva
vetustissima Europæ, Scytho-Celtica et Gothica . . . Wittenbergæ [Wittenberg]: typis
Christiani Schrödteri.
Kirchmaier, Georg Wilhelm [praeses] and Georg Friedrich Thryllitsch [respondens]
(1709). Eruditorum de dialecto Graecorum communi sententias . . . excutiendas
proponet . . . Thryllitius. VVitembergae [Wittenberg]: literis Christiani Schroedteri.
Klippi, Carita (2020). ‘Dialect and Its Technologies: Communities of Practice in Gallo-
Romance Dialectology’, in Van Rooy (2020b: 107–35).
Kloss, Heinz (1952). Die Entwicklung neuer germanischer Kultursprachen von 1800 bis
1950. Munich: Pohl.
Kloss, Heinz (1967). ‘ “Abstand Languages” and “Ausbau Languages” ’, Anthropological
Linguistics 9 (7): 29–41.
Kloss, Heinz (1978). Die Entwicklung neuer germanischer Kultursprachen seit 1800. 2.,
erweiterte Auflage. Sprache der Gegenwart 37. Düsseldorf: Pädagogischer Verlag
Schwann.
Koerner, E. F. K. (1989). ‘August Schleicher and Linguistic Science in the Second Half of the
19th Century’, in E. F. K. Koerner (ed.), Practicing Linguistic Historiography. Studies in
the History of the Language Sciences 50. Amsterdam & Philadelphia, PA: John
Benjamins, 324–75.
Kökeritz, Helge (1938). ‘Alexander Gill (1621) on the Dialects of South and East England’,
Studia Neophilologica 11 (1–2): 277–88.
Kramer, Johannes (1989). ‘Klassische Sprache und Substandard in der Geschichte des
Griechischen’, in Günter Holtus and Edgar Radtke (eds), Sprachlicher Substandard II:
Standard und Substandard in der Sprachgeschichte und in der Grammatik. Konzepte der
Sprach- und Literaturwissenschaft 44. Tübingen: Max Niemeyer, 55–82.
Kretzschmar, Jr, William A. (2009). The Linguistics of Speech. Cambridge & New York:
Kümin, Beat (2015). ‘The Uses of Space in Early Modern History—An Afterword’, in Paul
Stock (ed.), The Uses of Space in Early Modern History. Palgrave Studies in Cultural and
Intellectual History. New York: Palgrave Macmillan, 227–34.
Kurath, Hans (1972). Studies in Area Linguistics. Bloomington, IN & London: Indiana
University Press.
Labov, William (1963). ‘The Social Motivation of a Sound Change’, WORD 19 (3): 273–309.
Labov, William (1972). ‘The Social Stratification of (r) in New York City Department
Stores’, in William Labov, Sociolinguistic Patterns. Philadelphia, PA: University of
Pennsylvania Press, 43–54.
318 
Labov, William (1973). ‘Where Do Grammars Stop?’, in R. W. Shuy (ed.), Report on the
Twenty-Third Annual Round Table Meeting on Linguistics and Language Studies.
Georgetown Monograph Series on Language and Linguistics 25. Washington, DC:
Georgetown University Press, 43–88.
Labov, William (2010). Principles of Linguistic Change. Volume 3. Language in Society 39.
Malden, MA, Oxford & Chichester: Wiley-Blackwell.
Laes, Christian and Toon Van Houdt (2013). ‘Taalkunde met een ranzig randje:
Wetenschap en ideologie in het debat tussen Grotius en De Laet over de oorsprong en
de status van de Indianen (1641–1644)’, in Toon Van Hal et al. (eds), De tuin der talen:
Taalstudie en taalcultuur in de Lage Landen, 1450–1750. Orbis Linguarum: Studies op
het gebied van de taalbeschrijving, de taalfilosofie, en de taal- en cultuurgeschiedenis 3.
Leuven, Paris & Walpole, MA: Peeters, 125–52.
Lallot, Jean (1995). ‘Analogie et pathologie dans la grammaire alexandrine’, Lalies
15: 109–23.
Lambert, Frédéric (2009). ‘Les noms des langues chez les Grecs’, Histoire Épistémologie
Langage 31 (2): 15–27.
Lamers, Han (2015). Greece Reinvented: Transformations of Byzantine Hellenism in
Renaissance Italy. Brill’s Studies in Intellectual History 247. Leiden: Brill.
Lamers, Han, Toon Van Hal, and Sebastiaan G. Clercx (2020). ‘How to Deal with Scholarly
Forgetting in the History of the Humanities: Starting Points for Discussion’, History of
Humanities 5 (1): 5–29.
Lang, Jürgen (1982). Sprache im Raum: Zu den theoretischen Grundlagen der
Mundartforschung. Unter Berücksichtigung des Rätoromanischen und Leonesischen.
Beihefte zur Zeitschrift für romanische Philologie 185. Tübingen: Max Niemeyer.
Larramendi, Manuel de (1729). El impossible vencido: Arte de la lengua bascongada. En
Salamanca: por Antonio Joseph Villargordo Alcaráz.
Lascaris, Constantine (1502). De octo partibus orationis liber primus . . . [Venice]: [Aldus
Manutius].
Law, Vivien (2003). The History of Linguistics in Europe: From Plato to 1600. Cambridge:
Leerssen, Joep (2007). ‘Imagology: History and Method’, in Manfred Beller and Joep
Leerssen (eds), Imagology. The Cultural Construction and Literary Representation of
National Characters: A Critical Survey. Amsterdam & New York: Rodopi, 17–32.
Leibniz, Gottfried Wilhelm (1710). ‘Brevis designatio meditationum de originibus gentium,
ductis potissimum ex indicio linguarum’, Miscellanea Berolinensia . . . 1: 1–16.
Leskien, August (1876). Die Declination im Slavisch-litauischen und Germanischen.
Preisschriften gekrönt und herausgegeben von der fürstlich jablonowski’schen
Gesellschaft zu Leipzig. Nr. XI der historisch-nationalökonomischen Section 19.
Leipzig: Bei S. Hirzel.
Leskien, August (1879). ‘Giese, Albert’, in Allgemeine Deutsche Biographie. Leipzig:
Duncker & Humblot, 9: 151.
Leu, Urs B. and Peter Opitz, eds (2019). Conrad Gessner (1516–1565): Die Renaissance der
Wissenschaften/The Renaissance of Learning. Berlin & Boston, MA: De Gruyter
Oldenbourg.
Leu, Urs B., Raffael Keller, and Sandra Weidmann (2008). Conrad Gessner’s Private Library.
History of Science and Medicine Library 5. Leiden & Boston, MA: Brill.
Leusden, Johannes (1656). Philologus Hebræus . . . Ultrajecti [Utrecht]: apud Meinardum à
Dreunen.
 319
Leusden, Johannes (1670). ‘Dissertatio undecima, de dialectis Novi T. in genere & in specie’,
in Philologus Hebræo-Græcus generalis . . . Ultrajecti [Utrecht]: ex officinâ Anthonii
Smytegelt, 83–91.
Liddell, Henry George and Robert Scott (1996). A Greek–English Lexicon. Revised and
augmented by Sir Henry Stuart Jones with the assistance of Roderick McKenzie and with
the cooperation of many scholars. Oxford: Clarendon Press.
van der Linde, J. M. (1987). Jan Willem Kals: Leraar der hervormden; advocaat van indiaan
en neger. Kampen: J. H. Kok.
Löffler, Heinrich (1974). Probleme der Dialektologie: Eine Einführung. Darmstadt:
Wissenschaftliche Buchgesellschaft.
Lusignan, Serge (1986). Parler vulgairement: Les intellectuels et la langue française aux XIIIe
et XIVe siècles. Études médiévales. Paris & Montréal: Librairie philosophique J. Vrin &
Les Presses de l’Université de Montréal.
Luther, Martin (1913). Werke: Tischreden, 1531–1546, ed. Ernst Kroker. Volume 2.
D. Martin Luthers Werke: Kritische Gesamtausgabe 2. Weimar: Hermann Böhlaus
Nachfolger.
Luther, Martin (1916). Werke: Tischreden, 1531–1546, ed. Ernst Kroker. Volume 4.
D. Martin Luthers Werke: Kritische Gesamtausgabe 2. Weimar: Hermann Böhlaus
Nachfolger.
Luther, Martin (1919). Tischreden 1531–46, ed. Ernst Kroker. Volume 5. D. Martin Luthers
Werke: Kritische Gesamtausgabe 2. Weimar: Hermann Böhlaus Nachfolger.
Lynn, Michael R. (2006). Popular Science and Public Opinion in Eighteenth-Century France.
Studies in Early Modern European History. Manchester: Manchester University Press.
Macaulay, Ronald (2010). ‘Dialect’, in Mirjam Fried et al. (eds), Variation and Change:
Pragmatic Perspectives. Handbook of Pragmatics Highlights 6. Amsterdam &
Philadelphia, PA: John Benjamins, 61–72.
Machan, Tim William (2003). English in the Middle Ages. Oxford: Oxford University Press.
Maierù, Alfonso (1983). ‘Dante al crocevia?’, Studi medievali 24: 735–48.
Malkiel, Yakov (1967). ‘Uriel Weinreich (1926–1967)’, Language 43 (2): 605–7.
Manutius, Aldus (1496). ‘Aldus Manucius Basianâs Romanus studiosis omnibus .S .P. D.’,
in Aldus Manutius et al. (1496: *.ii–iii).
Manutius, Aldus et al., eds (1496). Θησαυρός. Κέρας ἀμαλθείας, καὶ κῆποι Ἀδώνιδος.
Thesaurus Cornu copiæ. & Horti Adonidis. Venetiis [Venice]: in domo Aldi Romani.
Marazzini, Claudio (1993). ‘Le teorie’, in Luca Serianni and Pietro Trifone (eds), Storia della
lingua italiana. Torino: Giulio Einaudi, 1: 231–329.
Marenbon, John (2004). ‘Mare, William de La [William de Mara] (Fl. 1272–1279),
Franciscan Friar and Theologian’, in Oxford Dictionary of National Biography. Online.
Available at www.oxforddnb.com/view/10.1093/ref:odnb/9780198614128.001.0001/odnb-
9780198614128-e-18025. Last accessed 14 January 2020.
Martinet, André (1954). ‘Dialect’, Romance Philology 8: 1–11.
Matheeussen, Constant (1974). Een zestiende-eeuws pleidooi tegen de grammatica: Georgius
Haloinus en zijn Restauratio linguae Latinae (1533). Leuven: unpublished PhD disser-
tation (KU Leuven).
Matheeussen, Constant (1977). ‘Haloinus (Halewijn, Halewyn, ook Cominius), Georgius’,
in Nationaal Biografisch Woordenboek. Brussel: Paleis der Academiën, 7: 310–17.
Mattheier, Klaus, ed. (1983). Aspekte der Dialekttheorie. Tübingen: Max Niemeyer.
Maxwell, Alexander (2018). ‘When Theory Is a Joke: The Weinreich Witticism in
Linguistics’, Beiträge zur Geschichte der Sprachwissenschaft 28 (2): 263–92.
320 
May, Stephen (2011). ‘Language Policy’, in Michael Grenfell, Bourdieu, Language and
Linguistics. London & New York: Continuum, 147–69.
Mayer, Bartholomaeus (1629). Philologiæ sacræ pars prima . . . Lipsiæ [Leipzig]: sumtibus
Gothofredi Grosii . . . Excudebat Johann-Albertus Minzelius.
Mayrhofer, Manfred (1983). Sanskrit und die Sprachen Alteuropas: Zwei Jahrhunderte des
Widerspiels von Entdeckungen und Irrtümern. Göttingen: Vandenhoeck & Ruprecht.
Mazzarella-Farao, Francesco (1779). La Neoellenopedia o sia il nuovo metodo per erudire la
gioventù nel greco linguaggio. Volume 1. Napoli [Naples]: nella stamperia porsiliana.
Mazzocco, Angelo (1993). Linguistic Theories in Dante and the Humanists: Studies of
Language and Intellectual History in Late Medieval and Early Renaissance Italy. Brill’s
Studies in Intellectual History 38. Leiden, New York & Cologne: Brill.
McCain, Stewart (2018). The Language Question under Napoleon. War, Culture and
Society, 1750–1850. Cham: Palgrave Macmillan.
McWhorter, John (2016). ‘What’s a Language, Anyway?’, The Atlantic, 19 January 2016.
Online. Available at www.theatlantic.com/international/archive/2016/01/difference-
between-language-dialect/424704/. Last accessed 14 January 2020.
McWhorter, John (2019). ‘Language vs. Dialect’, Lexicon Valley, 6 August 2019. Online.
Available at https://slate.com/podcasts/lexicon-valley/2019/08/john-mcwhorter-on-
language-versus-dialect>. Last accessed 14 January 2020.
Meillet, Antoine (1903). Introduction à l’étude comparative des langues indo-européennes.
Paris: Hachette.
Meillet, Antoine (1908). Les dialectes indo-européens. Collection linguistique publiée par la
Société de linguistique de Paris 1. Paris: Honoré Champion.
Meillet, Antoine (1924). ‘Introduction’, in Antoine Meillet and Marcel Cohen (eds), Les
langues du monde. Collection linguistique publiée par la Société de linguistique de Paris
16. Paris: Édouard Champion, 1–18.
Meillet, Antoine (1925). La méthode comparative en linguistique historique. Oslo, Leipzig,
Paris, London & Cambridge, MA: H. Aschehoug and Co. (W. Nygaard), Otto
Harrassowitz, Honoré Champion, Williams & Norgate & Harvard University Press.
Melanchthon, Philipp (1518). Institutiones Graecae grammaticae. Tubingæ & Hagnoæ
[Tübingen & Haguenau]: ex Academia Anshelmiana.
Merton, Robert K. (1968). Social Theory and Social Structure. Enlarged edition. New York:
The Free Press.
Merula, Paulus (1605). Cosmographiæ generalis libri tres . . . [Leiden]: ex Officina
Plantiniana Raphelengi F.
Metcalf, George J. (2013). On Language Diversity and Relationship from Bibliander to
Adelung, ed. Toon Van Hal and Raf Van Rooy. Studies in the History of the Language
Sciences 120. Amsterdam & Philadelphia, PA: John Benjamins.
Meyer, John W., Francisco O. Ramirez, and Yasemin Nuholu Soysal (1992). ‘World
Expansion of Mass Education, 1870–1980’, Sociology of Education 65 (2): 128–49.
Meyer-Lübke, Wilhelm (1908). ‘Roger Bacon über die französischen Mundarten’,
Zeitschrift für französische Sprache und Literatur 32: 175.
Michaeler, Karl Joseph (1776). Tabulae parallelae antiquissimarum Teutonicae linguae
dialectorum . . . Oeniponte [Innsbruck]: formis Academicis Wagnerianis.
Mickey, Katherine (1981). ‘Dialect Consciousness and Literary Language: An Example
from Ancient Greek’, Transactions of the Philological Society 79: 35–66.
Miechowita, Maciej (1517). Tractatus de duabus Sarmatijs Asiana: & Europiana & de
contentis in eis. Cracouię [Cracow]: impressum . . . opera & impensis prouidi viri
d[omi]ni Ioannis Haller.
 321
Mokyr, Joel (2002). The Gifts of Athena: Historical Origins of the Knowledge Economy.
Princeton, NJ: Princeton University Press.
Molland, George (2004). ‘Bacon, Roger (c.1214–1292?)’, in Oxford Dictionary of National
Biography. Online. Available at www.oxforddnb.com/view/article/1008. Last accessed 10
August 2020.
Moos, Peter von, ed. (2008). Zwischen Babel und Pfingsten: Sprachdifferenzen und
Gesprächsverständigung in der Vormoderne (8.–16. Jahrhundert) . . . Zürich & Berlin:
LIT.
Morhof, Daniel Georg (1685). De Patavinitate Liviana liber: Ubi de urbanitate & peregri-
nitate sermonis Latini universè agitur. Kiloni [Kiel]: literis ac sumptibus Joachimi
Reumanni.
Morpurgo Davies, Anna (1987). ‘The Greek Notion of Dialect’, Verbum 10: 7–28.
Morpurgo Davies, Anna (1998). History of Linguistics, ed. Giulio C. Lepschy. Volume 4.
Longman Linguistics Library. London & New York: Longman.
Mosellanus, Petrus (1527). In M[arci] Fab[ii] Quintiliani Rhetoricas institutiones annota-
tiones. Basileæ [Basel]: apud Adamum Petrum.
Muller, Jean-Claude (1984a). ‘Quelques repères pour l’histoire de la notion de vocabulaire
de base dans le précomparatisme’, Histoire Épistémologie Langage 6 (2): 37–43.
Muller, Jean-Claude (1984b). ‘Saumaise, Monboddo, Adelung: Vers la grammaire
comparée’, in Sylvain Auroux et al. (eds), Matériaux pour une histoire des théories
linguistiques . . . Travaux et recherches. Lille: Université de Lille III, diffusion P.U.L.,
389–96.
Muller, Jean-Claude (1986). ‘Early Stages of Language Comparison from Sassetti to Sir
William Jones (1786)’, Kratylos 31: 1–31.
Müller, Roman (2001). Sprachbewusstsein und Sprachvariation im lateinischen Schrifttum
der Antike. Munich: C. H. Beck.
Murray, Robert W. (2010). ‘Language and Space: The Neogrammarian Tradition’, in Auer
and Schmidt (2010: .70–87).
Mylius, Abraham (1612). Lingua Belgica . . . Lugduni Batavorum [Leiden]: pro Bibliopolio
Commeliniano, excudebant . . . Vlricus Cornelij & G. Abrahami.
Myricaeus, Johann Caspar (1619). Grammaticæ Syro-Chaldææ libri duo . . . [Geneva]: ex
typographεῖα Syra Petri de la Rouiere.
Nat, Jan (1929). De studie van de Oostersche talen in Nederland in de 18e en de 19e eeuw.
Purmerend: J. Muusses.
Neis, Cordula (2009). ‘Korruption’, in Haßler and Neis (2009: .567–82).
Newton, Brian (1972). The Generative Interpretation of Dialect: A Study of Modern Greek
Phonology. Cambridge Studies in Linguistics 8. Cambridge: at the University Press.
Nibbe, Johann Barthold (1725). Dialectologia paradigmatica oder Gründliche Anweisung zu
denen 4. Haupt-dialectis der Griechischen Sprache . . . Neu-Brandenburg: verlegts Joh. George
Schönborn.
Nicholas of Lyra (1471). [Postilla litteralis super Matthaeum et Ioannem]. [Rome]: [Konrad
Sweinheim & Arnold Pannartz].
Niebaum, Hermann and Jürgen Macha (2006). Einführung in die Dialektologie des
Deutschen. 2., neubearbeitete Auflage. Germanistische Arbeitshefte 37. Tübingen: Max
Niemeyer.
Noordegraaf, Jan (2014). ‘Verre verwanten: D. C. Hesseling (1859–1941) over taal en
taalverandering’, in Freek Van de Velde et al. (eds), Patroon en argument: Een dubbel-
feestbundel bij het emeritaat van William Van Belle en Joop van der Horst. Leuven:
Universitaire Pers, 523–35.
322 
Ogilvie, Brian W. (2006). The Science of Describing: Natural History in Renaissance Europe.
Chicago, IL & London: The University of Chicago Press.
Osthoff, Hermann and Karl Brugmann (1878). ‘Vorwort’, Morphologische Untersuchungen
auf dem Gebiete der indogermanischen Sprachen 1: –.
Oxford English Dictionary. Online. Available at www.oed.com/. Last accessed 10 August
2020.
Padley, G. A. (1988). Grammatical Theory in Western Europe 1500–1700: Trends in
Vernacular Grammar II. Cambridge: Cambridge University Press.
Papy, Jan, ed. (2018). The Leuven Collegium Trilingue 1517–1797: Erasmus, Humanist
Educational Practice and the New Language Institute Latin – Greek – Hebrew. Leuven,
Paris & Bristol, CT: Peeters.
Pascale, Carlo (1598). Legatus. Rothomagi [Rouen]: apud Raphaelem Parvivallium.
Pasor, Georg (1632). Syllabus brevis Grecarum Novi Testamenti dialectorum. Franekeræ
[Franeker]: excudebat Fredericus Heynsius.
Paul, Hermann (1880). Principien der Sprachgeschichte. Halle: Max Niemeyer.
Pennisi, Antonino (2009). ‘Galiani, Ferdinando’, in Stammerjohann (2009: .508).
Percival, W. Keith (1992). ‘La connaissance des langues du monde’, in Auroux (1992:
226–38).
Perotti, Niccolò (1489). Cornucopiæ seu come[n]tariorum [sic] linguæ Latinæ. Venetiis
[Venice]: per magistrum Paganinum de Paganinis Brixiensem.
Persons, Robert (1585). A Christian directorie guiding men to their salvation . . . [S. l.]: [s. n.].
Peters, Manfred (1970–1). Conrad Gessner als Germanist und Linguist. Four volumes.
Ghent: unpublished PhD dissertation (Rijksuniversiteit Gent).
Peters, Manfred (1972). ‘Sprachwandel und Sprachnorm in Conrad Gessners Mithridates
und in seiner Vorrede zu Josua Maalers Teütsch spraach’, Archiv für das Studium der
neueren Sprachen und Literaturen 208 (4–6): 256–66.
Peucer, Caspar (1562). Tertia pars Chronici Carionis, a Carolo Magno, ubi Philippus
Melanthon desijt, usq[ue] ad Fridericum Secundum. VVitebergae [Wittenberg]: excude-
bat Petrus Seitz.
Pfaffel, Wilhelm (1981). Quartus gradus etymologiae: Untersuchungen zur Etymologie
Varros in ‘De lingua Latina’. Beiträge zur klassischen Philologie 131. Königstein/Ts.
[Königstein im Taunus]: Anton Hain.
Pfeiffer, August [praeses] and Johann Georg Martini [respondens] (1663). Diatribe philolo-
gica de lingua Galilæa, per quam D. Petrus agnitus fuisse legitur . . . Wittebergæ
[Wittenberg]: typis Johannis Haken.
Pfeiffer, August [praeses] and Johann Georg Martini [respondens] (1699). ‘Exercitatio
decima de dialecto Galilæa, per quam D. Petrus agnitus fuisse legitur . . . ’, in August
Pfeiffer, Dubia vexata Script[urae] Sacræ . . . Editio quarta. Lipsiæ [Leipzig]: apud Joh.
Frider. Gleditschium, 206–16 [second sequence of pagination].
Phillips, Edward (1658). The new world of English words: or a general dictionary . . . London:
printed by E. Tyler, for Nath. Brooke.
Piccolomini, Enea Silvio (1960). Ausgewählte Texte aus seinen Schriften . . . , ed./transl.
Berthe Widmer. Basel & Stuttgart: Benno Schwabe.
Pierce, Joe E. (1952). ‘Dialect Distance Testing in Algonquian’, International Journal of
American Linguistics 18 (4): 203–10.
Pinborg, Jan (1967). Die Entwicklung der Sprachtheorie im Mittelalter. Münster:
Aschendorff.
Pirazzini, Daniela (2013). Theorien und Methoden der romanischen Sprachwissenschaft.
Romanistische Arbeitshefte 59. Berlin & Boston, MA: Walter de Gruyter.
 323
Polák, Václav (1954). ‘Contribution à l’étude de la notion de langue et de dialecte’, Orbis:

Bulletin international de documentation linguistique 3 (1): 89–98.
Polier de Bottens, Antoine-Noé [auctor et defendens] (1739). Dissertatio philologica qua
disquiritur de puritate dialecti Arabicæ, comparate cum puritate dialecti Hebrææ, in
relatione ad antediluvianam linguam . . . Lugduni in Batavis [Leiden]: apud Johannem
Luzac.
Poole, Matthew (1683). Annotations upon the Holy Bible . . . Volume 1. London: printed by
John Richardson, for Thomas Parkhurst . . .
Pop, Sever (1950). La dialectologie: Aperçu historique et méthodes d’enquêtes linguistiques.
Two volumes. Université de Louvain: Recueil de travaux d’histoire et de philologie 3e
série 38. Louvain & Gembloux: chez l’auteur & J. Duculot.
Poser, William J. and Lyle Campbell (1992). ‘Indo-European Practice and Historical
Methodology’, in Laura A. Buszard-Welcher et al. (eds), Proceedings of the Eighteenth
Annual Meeting of the Berkeley Linguistics Society (February 14–17, 1992): General
Session and Parasession on the Place of Morphology in a Grammar. Berkeley, CA:
Berkeley Linguistic Society, 214–36.
Pott, August Friedrich (1856). Die Ungleichheit menschlicher Rassen hauptsächlich
vom sprachwissenschaftlichen Standpunkte, unter besonderer Berücksichtigung von des
Grafen von Gobineau gleichnamigem Werke . . . Lemgo & Detmold: Meyer’sche
Hofbuchhandlung.
Pott, August Friedrich (1974 [1884–90]). Einleitung in die Allgemeine Sprachwissenschaft
(1884–1890) together with Zur Literatur der Sprachenkunde Europas (Leipzig, 1887), ed.
E. F. K. Koerner. Amsterdam Classics in Linguistics, 1800–1925 10. Amsterdam: John
Benjamins.
Pozzi, Mario (1978). Trattatisti del Cinquecento. Milan: Riccardo Ricciardi.
Püschel, Ulrich (2009). ‘Fulda, Friedrich Carl’, in Stammerjohann (2009: .498).
Puttenham, George (1589). The arte of English poesie . . . At London: printed by Richard Field.
Pyle, Cynthia M. (2000). ‘Conrad Gessner on the Spelling of His Name’, Archives of Natural
History 27 (2): 175–86.
R. L. G. (2014). ‘How a Dialect Differs from a Language: The Economist Explains’. The
Economist (Online), 16 February 2014. Online. Available at www.economist.com/the-
economist-explains/2014/02/16/how-a-dialect-differs-from-a-language. Last accessed 10
August 2020.
Ran, Qibin and Søren Wichmann (2018). ‘Zenyang qufen yuyan yu fangyan: Jiyu hexin
cihui de juli jisuan fangfa tansuo [How to Distinguish Languages and Dialects: A
Quantitative Approach]’, Yuyan Zhanlue Yanjiu [Chinese Journal of Language Policy
and Planning] 3 (2): 50–8.
Raphelengius, Franciscus (1983). ‘349. 84 05 18 R. Franciscus Raphelengius sr. (Antwerpen)
aan Lipsius (Leiden)’, in M. A. Nauwelaerts and S. Sué (eds), Iusti Lipsi Epistolae 2.
Brussel: Paleis der Academiën, 121–3.
Rask, Rasmus (2013 [1818]). Investigation of the Origin of the Old Norse or Icelandic
Language. New edition of the 1993 English translation by Niels Ege, with an introduction
by Frans Gregersen. Amsterdam Classics in Linguistics, 1800–1925 18. Amsterdam &
Philadelphia, PA: John Benjamins.
Rastell, John (1566). A treatise intitled, beware of M. Iewel. Antuerpiæ [Antwerp]: ex
officina Ioannis Fouleri.
Ravis, Christian (1646). Orthographiæ et analogiæ (vulgò etymologiæ) Ebraicæ delineatio
juxta vocis partes abstractas I. consonas. II. vocales. III. Accentus . . . Amstelodami
[Amsterdam]: apud Joannem Janssonium.
324 
Ravis, Christian (1650). A generall grammer for the ready attaining of the Ebrew, Samaritan,
Calde, Syriac, Arabic, and the Ethiopic languages . . . London: printed by W. Wilson for
Tho. Slater and Tho. Humington.
Ravis, Christian [praeses et auctor] (1656). Disputatio de Dudaim Rubenis . . . Ubsaliæ
[Uppsala]: excudebat Johannes Pauli.
Ray, John (1678). A collection of English proverbs . . . The second edition enlarged by the
addition of many hundred English . . . Cambridge: printed by John Hayes . . . for
W. Morden.
Regis, Riccardo (2012). ‘Note su koinè’, Romanische Forschungen 124 (1): 3–16.
Rhellicanus, Johannes (1543). In C. Iulii Cæsaris dictatoris uiri disertissimi, et Auli Hirtij,
seu Oppij, Commentaria de bello Gallico, ciuili Pompeiano, Alexandrino, Africano, &
Hispaniensi, annotationes . . . Basileæ [Basel]: per Hieronymum Curionem.
Ribier, Guillaume (1666). ‘Remarques sur ladite ville de Soleurre et sur la diversité des
langues d’Allemagne’, in Guillaume Ribier (ed.), Lettres et mémoires d’estat des roys,
princes, ambassadeurs et autres ministres, sous les regnes de François premier, Henry II. et
François II. . . . A Paris: chez François Clouzier . . . et la vefve Aubouyn, 1: 25–8.
Rice, Eugene F. and Anthony Grafton (1994). The Foundations of Early Modern Europe,
1460–1559. Second edition. New York: Norton.
Richey, Michael (1743). Idioticon Hamburgense sive Glossarium vocum Saxonicarum quae
populari nostra dialecto Hamburgi maxime frequentantur . . . Hamburgi [Hamburg]:
apud Conr. Koenigium.
Richey, Michael (1755). Idioticon Hamburgense oder Wörter-Buch, zur Erklärung der
eigenen, in und üm Hamburg gebräuchlichen, Nieder-Sächsischen Mund-Art. Jetzo
vielfältig vermehret . . . Hamburg: verlegt von Conrad König.
Rider, John (1589). Bibliotheca scholastica: A double dictionarie . . . Oxford: printed by
Joseph Barnes.
Rijcke, Theodorus (1684). ‘Dissertatio de primis Italiæ colonis et Æneæ adventu’, in Lucas
Holstenius, Notæ et castigationes posthumæ in Stephani Byzantii Ἐθνικά, Quæ vulgo
Περὶ πόλεων inscribuntur . . . , ed. Theodorus Rijcke. Lugd[uni] Batavorum [Leiden]:
Apud Jacobum Hackium, 399–467.
Rizzo, Silvia (1990). ‘Petrarca, il latino e il volgare’, Quaderni petrarcheschi 7: 7–40.
Robins, R. H. (1967). A Short History of Linguistics. Longman Linguistics Library 6.
London: Longman.
Rocher, Rosane (2001). ‘The Knowledge of Sanskrit in Europe until 1800’, in Auroux et al.
(2001: 1156–63).
Roest, Bert (2015). Franciscan Learning, Preaching and Mission c. 1220–1650: Cum scientia
sit donum Dei, armatura ad defendendam sanctam Fidem catholicam . . . Leiden &
Boston, MA: Brill.
Rollin, Charles (1726). De la maniere d’enseigner et etudier les belles lettres, par raport à
l’esprit & au cœur. Volume 1. A Paris: chez Jacques Estienne.
Ronsard, Pierre de (1565). Abbregé de l’art poëtique françois. A Paris: chez Gabriel Buon.
Rosier-Catach, Irène (1997). ‘Roger Bacon and Grammar’, in Jeremiah Hackett (ed.), Roger
Bacon and the Sciences: Commemorative Essays. Studien und Texte zur Geistesgeschichte
des Mittelalters 57. Leiden, New York & Cologne: Brill, 67–102.
Rotolo, Vincenzo (1973–4). ‘L’opinione di F. Filelfo sul greco volgare’, Rivista di studi
bizantini e neoellenici 20–1 (n. s. 10–11): 85–107.
Rousseau, Jean-Jacques (1781 [1755]). ‘Essai sur l’origine des langues, où il est parlé de
la mélodie & de l’imitation musicale’, in Œuvres posthumes de Jean-Jaques
Rousseau . . . Volume 3. Genève [Geneva]: [s. n.], 3: 211–327.
 325
Rubin, Donald L. (1992). ‘Nonlanguage Factors Affecting Undergraduates’ Judgments of

Nonnative English-Speaking Teaching Assistants’, Research in Higher Education 33 (4):
511–31.
Ruddiman, Walter, ed. (1783). ‘To query XV.–“What are the causes of a difference in
dialect among people who speak the same language?”’, The Edinburgh weekly magazine
57 (September): 334.
Ruhig, Philipp (1745). Betrachtung der littauischen Sprache in ihrem Ursprunge, Wesen und
Eigenschaften . . . Königsberg [modern Kaliningrad]: druckts und verlegts Johann
Heinrich Hartung.
Sabellicus, Marcus Antonius (1490). Suetonius cum commento. Venetiis [Venice]: per
Baptistam de Tortis.
Sacré, Dirk (1994). ‘Andreas Hoius’ sterfjaar (met een noot over zijn familie)’, Biekorf:
Westvlaams archief voor geschiedenis, archeologie, taal- en volkskunde 94: 387–92.
Sansone, Mario (1999). Da Bembo a Galiani: Il dibattito sulla lingua in Italia, ed. Francesco
Tateo. Biblioteca di Filologia Romanza 41. Bari: Adriatica.
Sapir, Edward (1921). Language: An Introduction to the Study of Speech. New York:
Harcourt, Brace and Company.
Sapir, Edward (1931). ‘Dialect’, in Encyclopaedia of the Social Sciences. New York:
Macmillan, 5: 123–6.
Saporta, Sol (1965). ‘Ordered Rules, Dialect Differences, and Historical Processes’,
Language 41 (2): 218–24.
Saumaise, Claude de (1643a). De Hellenistica commentarius, controversiam de lingua
Hellenistica decidens, & plenissimè pertractans originem et dialectos Græcæ linguæ.
Lugd[uni] Batavor[um] [Leiden]: ex Officinâ Elseviriorum.
Saumaise, Claude de (1643b). Funus linguæ Hellenisticæ, siue Confutatio exercitationis de
Hellenistis et lingua Hellenistica. Lugduni Batavorum [Leiden]: ex Officina Ioannis
Maire.
Saussure, Ferdinand de (1995 [1916]). Cours de linguistique générale, ed. Charles Bally and
Albert Sechehaye. Édition critique préparée par Tullio de Mauro. Postface de Louis-Jean
Calvet. Paris: Payot.
Scaliger, Joseph Justus (1594). Epistola de vetustate et splendore gentis Scaligeræ, et Iul[ii]
Cæs[aris] Scaligeri vita. Lugduni Batavorum [Leiden]: ex officina Plantiniana.
Scaliger, Joseph Justus (1610a). ‘Diatriba de Europæorum linguis’, in Isaac Casaubon (ed.),
Opuscula varia antehac non edita. Parisiis [Paris]: apud Hadrianum Beys, 119–22.
Scaliger, Joseph Justus (1610b). ‘Diatriba de hodiernis Francorum linguis’, in Isaac
Casaubon (ed.), Opuscula varia antehac non edita. Parisiis [Paris]: apud Hadrianum
Beys, 123–6.
Schäfer, Gottfried Heinrich (1811). Gregorii Corinthii et aliorum grammaticorum libri de
dialectis linguae Graecae . . . Lipsiae [Leipzig]: sumtibus Io. Aug. Gottl. Weigel.
Scheuerer, Franz Xaver (1995). Zum philologischen Werk J. A. Schmellers und seiner
wissenschaftlichen Rezeption: Eine Studie zur Wissenschaftsgeschichte der Germanistik.
Studia Linguistica Germanica 37. Berlin & New York: Walter de Gruyter.
Schilter, Johann (1728). Thesaurus antiquitatum Teutonicarum, ecclesiasticarum, civilium,
literariarum. Præfationem præmisit Joannes Frickius. Volume 3. Ulmæ [Ulm]: sumpti-
bus Danielis Bartholomæi & filii.
Schleicher, August (1861). Compendium der vergleichenden grammatik der indogerma-
nischen sprachen. Two volumes. Weimar: Hermann Böhlau.
Schleicher, August (1863). Die Darwinsche Theorie und die Sprachwissenschaft. Weimar:
Hermann Böhlau.
326 
Schlieben-Lange, Brigitte (1992). ‘Reichtum, Energie, Klarheit und Harmonie: Die

Bewertung der Sprachen in Begriffen der Rhetorik’, in Susanne R. Anschütz (ed.),
Texte, Sätze, Wörter und Moneme: Festschrift für Klaus Heger zum 65. Geburtstag.
Heidelberg: Heidelberger Orientverlag, 571–86.
Schmeller, Johann Andreas (1821). Die Mundarten Bayerns grammatisch dargestellt.
Munich: Karl Thienemann.
Schmidt, Erasmus (1604). Tractatus de dialectis Græcorum principalibus, quæ sunt in parte
λέξεως . . . VVittebergæ [Wittenberg]: imprimebat Laurentius Seuberlich. Impensis
Samuel Selfisch.
Schmidt, Erasmus (1615). ‘Discursus de pronunciatione Græca, contra νεόϕυτον’, in
Erasmus Schmidt (ed.), Cyrilli vel ut alii volunt Johannis Philoponi opusculum
utilißimum de differentiis vocum Græcarum quoad tonum, spiritum, genus, &c.
Witebergæ [Wittenberg]: typis Richterianis, sumptibus B. Corvini, 213–55.
Schmidt, Johannes (1872). Die verwandtschaftsverhältnisse der indogermanischen sprachen.
Weimar: Hermann Böhlau.
Schmitt, Charles B. (1983). Aristotle and the Renaissance. Cambridge, MA: Harvard
University Press.
Schnell, Rüdiger (2008). ‘Vom Nicht- und Missverstehen im Mittelalter: Zu Hindernissen
sprachlicher und nichtsprachlicher Interaktion im Mittelalter’, in Moos (2008: 567–90).
Schoppe, Caspar (1636). Consultationes de scholarum & studiorum ratione deque prudentiæ
& eloquentiæ parandæ modis. Patavii [Padua]: apud Paulum Frambottum.
Schottel, Justus Georg (1663). Ausführliche Arbeit von der teutschen Haubt
Sprache . . . Braunschweig: gedrukt und verlegt durch Christoff Friederich Zilligern.
Schrambke, Renate (2010). ‘Language and Space: Traditional Dialect Geography’, in Auer
and Schmidt (2010: .87–107).
Schrieckius, Adrianus (1615). Monitorum secundorum libri V . . . Ypris Flandrorum
[Ypres]: ex Officina typographica Francisci Belletti.
Schuchardt, Hugo (1900). Über die Klassifikation der romanischen Mundarten: Probe-
Vorlesung gehalten zu Leipzig am 30. April 1870. Graz: [s. n.].
Schuchardt, Hugo (1922). Zur Kenntnis des Baskischen von Sara (Labourd). Abhandlungen
der Preussischen Akademie der Wissenschaften. Jahrgang 1922: Philosophisch-
historische Klasse 1. Berlin: Verlag der Akademie der Wissenschaften.
Schuchardt, Hugo (1928). Hugo Schuchardt-Brevier: Ein Vademecum der allgemeinen
Sprachwissenschaft, ed. Leo Spitzer. Zweite erweiterte Auflage. Mit einem Porträt Hugo
Schuchardts. Halle (Saale): Max Niemeyer.
Schultens, Albert (1732). Oratio altera de linguæ Arabicæ antiquissima origine, intima ac
sororia cum lingua Hebræa cognatione, nullisque seculis præflorata puritate . . . Lugduni
Batavorum [Leiden]: apud Samuelem Luchtmans.
Schultens, Albert (1737). Institutiones ad fundamenta linguæ Hebrææ . . . Lugduni
Batavorum [Leiden]: apud Johannem Luzac.
Schultens, Albert (1738a [1729]). ‘Oratio prima de linguæ Arabicæ antiquissima origine,
intima ac sororia cum lingua Hebræa affinitate, nullisque seculis præflorata puritate . . . ’,
in Schultens (1738b: separate pagination).
Schultens, Albert (1738b). Originum Hebræarum tomus secundus . . . Lugduni Batavorum
[Leiden]: apud Samuelem Luchtmans.
Schultens, Albert (1738c). Vetus et regia via Hebraïzandi, asserta contra novam et meta-
physicam hodiernam. Lugduni Batavorum [Leiden]: apud Joannem Luzac.
 327
Schultens, Albert (1739). Excursus tertius ad editionem primam et secundam dissertationis

de lingua primæva, ejusque additamentum apologeticum. Lugduni Batavorum [Leiden]:
apud Johannem Luzac.
Schultens, Albert (1748). ‘Præfatio’, in Albert Schultens (ed.), Proverbia Salomonis . . .
Lugduni Batavorum [Leiden]: apud Johannem Luzac, –.
Schwartz, Christian Gottlieb [praeses] and Abraham Helm [respondens] (1702). Disputatio
de causis dialectorum, speciatim Graecarum . . . Vitembergae [Wittenberg]: prelo
Gerdesiano.
Sechehaye, Albert (1908). Programme et méthodes de la linguistique théorique: Psychologie
du langage. Paris, Leipzig & Geneva: Honoré Champion, Otto Harrassowitz &
A. Eggimann.
Séguy, Jean (1973). ‘La dialectométrie dans l’Atlas linguistique de la Gascogne’, Revue de
linguistique romane 37: 1–24.
Sergeev, Michail (2019). ‘Der Mithridates (1555) zwischen Sprachmuseum und neulatei-
nischem Onomastikon: Einige Überlegungen zur Konzeption und zum Genre des
Gessnerschen Handbuchs’, in Leu and Opitz (2019: 517–32).
Sèrio, Luigi (1780). Lo vernacchio: Resposta a lo dialetto napoletano. Napole [Naples]:
[s. n.].
Shapiro, Michael C. and Harold F. Schiffman (1981). Language and Society in South Asia.
Columbia, MO: South Asia Books.
Shelford, Robert (1635). Five pious and learned discourses . . . Cambridge: printed by the
printers to the Universitie of Cambridge.
Siebenborn, Elmar (1976). Die Lehre von der Sprachrichtigkeit und ihren Kriterien: Studien
zur antiken normativen Grammatik. Amsterdam: B. R. Grüner.
Silverstein, Michael (1971). ‘Whitney on Language’, in Michael Silverstein (ed.), Whitney
on Language. Cambridge, MA & London: MIT Press, x–xxiii.
Simon, Fabien Dimitri (2011). Sortir de Babel: Une République des Langues en quête d’une
«langue universelle» à la Renaissance et à l’Âge classique? Rennes: unpublished PhD
dissertation (Université Rennes 2).
Skinner, Quentin (1969). ‘Meaning and Understanding in the History of Ideas’, History and
Theory 8 (1): 3–53.
Springer, Katharina B. and Ragnar K. Kinzelbach (2008). Das Vogelbuch von Conrad Gessner
(1516–1565): Ein Archiv für avifaunistische Daten. Berlin & Heidelberg: Springer.
Stachowski, Marek (2013). ‘Miechowita’s Knowledge of East European Languages (Mainly
Hungarian, Lithuanian and Tatar) Based on His Tractatus de duabus Sarmatiis (1517)’,
Studia Linguistica Universitatis Iagellonicae Cracoviensis 130: 309–16.
Stammerjohann, Harro, ed. (2009). Lexicon Grammaticorum: A Bio-Bibliographical
Companion to the History of Linguistics. Second edition, revised and enlarged.
Tübingen: Niemeyer.
Stanihurst, Richard (1577). ‘A treatise contayning a playne and perfect description of
Irelande, with an introduction . . . ’, in Raphael Holinshed, The historie of Irelande from
the first inhabitation thereof, unto the yeare 1509 . . . At London: imprinted for John
Hunne, 2: 1–28.
Stanton, Domna C. and Rebecca M. Wilkin (2010). ‘Volume Editors’ Introduction’, in
Gabrielle Suchon, A Woman Who Defends All the Persons of Her Sex: Selected
Philosophical and Moral Writings. The Other Voice in Early Modern Europe. Chicago,
IL & London: The University of Chicago Press, 1–51.
328 
Stapleton, Thomas (1566). A returne of untruthes upon M. Jewelles replie. In Antwerpe

[Antwerp]: printed . . . by John Latius.
Stevens, Benjamin (2006–7). ‘Aeolism: Latin as a Dialect of Greek’, The Classical Journal
102 (2): 115–44.
Stevin, Simon (1586). ‘Uytspraeck vande weerdicheyt der Duytschen tael’, in Simon Stevin,
De beghinselen der weeghconst. Tot Leyden [Leiden]: inde druckerye van Christoffel
Plantijn, by Françoys van Raphelingen, b.1–d.4.
Stiernhielm, Georg (1669). Babel destructa, seu Runa Suethica . . . Stockholmiæ
[Stockholm]: [s. n.].
Stiernhielm, Georg (1671). ‘De linguarum origine præfatio’, in Georg Stiernhielm (ed.),
D. N. Jesu Christi SS. Evangelia ab Ulfila Gothorum in Moesia episcopo circa annum à
nato Christo CCCLX. ex Græco Gothicè translata . . . Stockholmiæ [Stockholm]: typis
Nicolai Wankif, a.3–f.4.
Stiernhielm, Georg (1937). ‘Till David Blondel (Hösten 1651.)’, in Per Wieselgren (ed.),
Samlade skrifter av Georg Stiernhielm. Svenska författare utgivna av svenska vitterhets-
samfundet 8. Stockholm: Albert Bonniers, 3.1: 136–41.
Stockhammer, Robert (2014). Grammatik: Wissen und Macht in der Geschichte einer
sprachlichen Institution. Berlin: Suhrkamp.
Strodtmann, Johann Christoph (1756). Idioticon Osnabrugense. Leipzig und Altona
[Leipzig & Altona, Hamburg]: in der Kortenschen Buchhandlung.
Swadesh, Morris (1950). ‘Salish Internal Relationships’, International Journal of American
Linguistics 16 (4): 157–67.
Swadesh, Morris (1952). ‘Lexico-Statistic Dating of Prehistoric Ethnic Contacts: With
Special Reference to North American Indians and Eskimos’, Proceedings of the
American Philosophical Society 96 (4): 452–63.
Swadesh, Morris (1955). ‘Towards Greater Accuracy in Lexicostatistic Dating’,
International Journal of American Linguistics 21 (2): 121–37.
Swartling, Birger (1909). Georg Stiernhielm: Hans lif och verksamhet. Uppsala universitets
årsskrift: Filosofi, sprakvetenskap och historiska vetenskaper 4. Uppsala: Akademiska
bokhandeln.
Swiggers, Pierre (1981). ‘The Word Šibbōleṯ in Jud. .6.’, Journal of Semitic Studies 26 (2):
205–7.
Swiggers, Pierre (1984). ‘Adrianus Schrieckius: De la langue des Scythes à l’Europe linguis-
tique’, Histoire Épistémologie Langage 6 (2): 17–35.
Swiggers, Pierre (1997). Histoire de la pensée linguistique. Paris: Presses Universitaires de France.
Swiggers, Pierre (2000). ‘La canonisation d’un franc-tireur: Le cas de Hugo Schuchardt’, in
Wolfgang Dahmen et al. (eds), Kanonbildung in der Romanistik und in den
Nachbardisziplinen: Romanistisches Kolloquium XIV. Tübinger Beiträge zur Linguistik
449. Tübingen: Gunter Narr, 269–304.
Swiggers, Pierre (2009). ‘Le français et l’italien en lice: L’examen comparatif de leurs
qualités chez Henri Estienne’, Synergies Italie 5: 69–76.
Swiggers, Pierre (2017). ‘Indo-European Linguistics in the 19th and 20th Centuries:
Beginnings, Establishment, Remodeling, Refinement, and Extension(s)’, in Jared Klein
et al. (eds), Handbook of Comparative and Historical Indo-European Linguistics.
Handbooks of Linguistics and Communication Science 41.1. Berlin & Boston, MA: De
Gruyter Mouton, 171–210.
Tavoni, Mirko (1982). ‘The 15th-Century Controversy on the Language Spoken by the
Ancient Romans: An Inquiry into Italian Humanist Concepts of “Latin”, “Grammar”,
and “Vernacular”’, Historiographia Linguistica 9 (3): 237–64.
 329
Tavoni, Mirko (1984). Latino, grammatica, volgare: Storia di una questione umanistica.
Padova: Antenore.
Tavoni, Mirko (1986). ‘On the Renaissance Idea that Latin Derives from Greek’, Annali
della Scuola normale superiore di Pisa: Classe di lettere e filosofia 16 (1): 205–38.
Tavoni, Mirko (1987). ‘Contributo all’interpretazione di De vulgari eloquentia I 1–9’,
Rivista di letteratura italiana 5 (3): 385–453.
Tavoni, Mirko (1990). ‘“Ydioma Tripharium” (Dante, De Vulgari Eloquentia, I 8–9)’, in
Hans-Josef Niederehe and E. F. K. Koerner (eds), History and Historiography of
Linguistics: Papers from the Fourth International Conference on the History of the
Language Sciences (ICHoLS IV), Trier, 24 – 28 August 1987. Studies in the History of
the Language Sciences 51. Amsterdam & Philadelphia, PA: John Benjamins, 1: 233–47.
Tavoni, Mirko (1998). ‘1 Renaissance Linguistics: 1.2 Western Europe’, in Giulio Lepschy
(ed.), History of Linguistics. Longman Linguistics Library. London & New York:
Longman, 3: 2–108.
Taylor, Daniel J. (1996). De lingua Latina X: A New Critical Text and English Translation
with Prolegomena and Commentary. Studies in the History of the Language Sciences 85.
Amsterdam & Philadelphia, PA: John Benjamins.
Thomas, Jacques (1953). ‘Dialecte et patois: Esquisse d’une étude sémantique’, Romanica
Gandensia 1: 93–117.
Thryllitsch, Georg Friedrich [auctor et respondens] (1709). Suspiciones quasdam historico-
technicas de dialectis Græcis ex consideratione originum migrationumq[ue] Graecarum
nationum collectas . . . eruditorum examini modeste subiicio . . . Thryllitius . . . Vitembergae
[Wittenberg]: litteris Schroedterianis.
Tillinger, Gábor (2013). ‘Langues, dialectes et patois—Problèmes de terminologie dialecto-
logique: Réflexions sur la situation géolinguistique en France et la terminologie
française’, Argumentum 9: 1–18.
Tory, Geoffroy (1529). Champ fleury . . . A Paris: par Maistre Geofroy Tory de Bourges . . . et
par Giles Gourmont.
Trabant, Jürgen (2006). Europäisches Sprachdenken: Von Platon bis Wittgenstein. Munich:
C. H. Beck.
Trautmann, Thomas R. (1998). ‘The Lives of Sir William Jones’, in Alexander Murray (ed.),
Sir William Jones, 1746–1794: A Commemoration. Oxford: Oxford University Press,
91–121.
Tribulato, Olga (2014). ‘Dialectology (diálektos), Ancient Theories of ’, in Georgios
K. Giannakis (ed.), Encyclopedia of Ancient Greek Language and Linguistics. Leiden &
Boston, MA: Brill, 1: 457–61.
Trissino, Gian Giorgio (1529). Dialωgω . . . intitulatω il castellanω, nel quale si tratta de la
lingua italiana. [Vincenza]: [Tolomeo Ianiculo].
Trovato, Paolo (1984). ‘“Dialetto” e sinonimi (“idioma”, “proprietà”, “lingua”) nella termi-
nologia linguistica quattro- e cinquecentesca (con un’appendice sulla tradizione a
stampa dei trattatelli dialettologici bizantini)’, Rivista di letteratura italiana 2: 205–36.
Trudeau, Danielle (1992). Les inventeurs du bon usage (1529–1647). Arguments. Paris:
Editions de Minuit.
Trudgill, Peter (1983). On Dialect: Social and Geographical Perspectives. Oxford: Basil
Blackwell.
Trudgill, Peter (2003). ‘Modern Greek Dialects: A Preliminary Classification’, Journal of
Greek Linguistics 4: 45–64.
Ursin, Georg Heinrich (1691). Grammatica Græca . . . Norimbergæ [Nuremberg]: sumpti-
bus Wolfgangi Mauritii Endteri.
330 
Vandewalle, Carolus Borromaeus (1929). ‘Roger Bacon dans l’histoire de la philologie’, La

France franciscaine 11 and 12: 315–410 and 45–90, 161–210.
Van Hal, Toon (2010a). ‘Moedertalen en taalmoeders’: Het vroegmoderne taalvergelijkende
onderzoek in de Lage Landen. Verhandelingen van de Koninklijke Vlaamse Academie
van België voor Wetenschappen en Kunsten. Nieuwe Reeks 20. Brussels: Koninklijke
Vlaamse Academie voor Wetenschappen en Kunsten.
Van Hal, Toon (2010b). ‘“Quam enim periculosa sit ea via . . . ”: Josephus Justus Scaliger’s
Views on Linguistic Kinship’, Beiträge zur Geschichte der Sprachwissenschaft 20 (1):
111–40.
Van Hal, Toon (2013/14). ‘From Alauda to Zythus: The Emergence and Uses of Old-
Gaulish Word Lists in Early Modern Publications’, Keltische Forschungen 6: 219–77.
Van Hal, Toon (2015). ‘Friedrich Gedike on Why and How to Compare the World’s
Languages: A Stepping Stone between Gottfried Wilhelm Leibniz and Wilhelm von
Humboldt?’, Beiträge zur Geschichte der Sprachwissenschaft 25 (1): 53–76.
Van Hal, Toon (2018). ‘The Alleged Persian-Germanic Connection: A Remarkable Chapter
in the Study of Persian from the Sixteenth through the Nineteenth Centuries’, in Alireza
Korangy and Corey Miller (eds), Trends in Iranian and Persian Linguistics. Berlin &
Boston, MA: De Gruyter Mouton.
Van Hal, Toon (forthcoming). ‘Ludolf ’s Language Laws: Pitfalls in Documenting and
Comparing the World’s Languages’.
Van Hal, Toon and Raf Van Rooy (2017). ‘ “Differing Only in Dialect”, or How
Collocations Can Co-Shape Concepts’, Language & Communication 56: 95–109.
Van Rooy, Raf (2013). ‘ “Πόθεν οὖν ἡ τοσαύτη διαϕωνία;”: Greek Patristic Authors
Discussing Linguistic Origin, Diversity, Change and Kinship’, Beiträge zur Geschichte
der Sprachwissenschaft 23 (1): 21–54.
Van Rooy, Raf (2016a). ‘Struggling to Order Diversity: The Variegated Classifications of
Greek Dialects before the Rise of Modern Linguistics’, Studies in Greek Linguistics 36:
465–73.
Van Rooy, Raf (2016b). ‘The Ancient Greek and Byzantine Notion of Dialektos: A Criterial
Approach’, in Carlos Assunção et al. (eds), History of Linguistics 2014: Selected Papers
from the 13th International Conference on the History of the Language Sciences (ICHoLS
XIII), Vila Real, Portugal, 25–29 August 2014. Studies in the History of the Language
Sciences 126. Amsterdam & Philadelphia, PA: John Benjamins, 55–67.
Van Rooy, Raf (2016c). ‘The Diversity of Ancient Greek through the Eyes of a Forgotten
Grammarian: Petrus Antesignanus (ca. 1524/1525–1561) on the Notion of «Dialect»’,
Histoire Épistémologie Langage 38 (1): 123–40.
Van Rooy, Raf (2016d). ‘ “What Is a ‘Dialect’?”: Some New Perspectives on the History of
the Term διάλεκτος and Its Interpretations in Ancient Greece and Byzantium’, Glotta 92:
244–79.
Van Rooy, Raf (2018a). ‘Latin as a Variable Language: Livy’s Patauinitas through Early
Modern Eyes’, in Astrid Steiner-Weber and Franz Römer (eds), Acta Conventus Neo-
Latini Vindobonensis: Proceedings of the Sixteenth International Congress of Neo-Latin
Studies (Vienna 2015). Leiden & Boston, MA: Brill, 767–77.
Van Rooy, Raf (2018b). ‘Regional Language Variation in European Thought before 1500:
A Historical Sketch Reflecting Current Research’, Beiträge zur Geschichte der
Sprachwissenschaft 28 (2): 167–216.
Van Rooy, Raf (2018c). ‘«Plutarque dialectologue»: La pseudo-autorité de Plutarque dans le
discours sur les dialectes grecs à la Renaissance (ca. 1400–1670)’, Revue belge de
philologie et d’histoire 96: 1115–34.
 331
Van Rooy, Raf (2019). ‘Διάλεκτος, Dialectus, Dialect: A Word’s Curious Journey from
Ancient Greek to (Neo-)Latin and Beyond’, Latomus: Revue d’études latines 78: 733–70.
Van Rooy, Raf (2020a). Greece’s Labyrinth of Language: A Study in the Early Modern
Discovery of Dialect Diversity. History and Philosophy of the Language Sciences 2. Berlin:
Language Science Press.
Van Rooy, Raf, ed. (2020b). Essays in the History of Dialect Studies: From Ancient Greece to
Modern Dialectology. Münster: Nodus.
Van Rooy, Raf (2020c). ‘Tradition, Synthesis, and Innovation: An Early Eighteenth-
Century Dissertation on Dialects Presented in Wittenberg’, in Marti Hanspeter and
Robert Seidel (eds), Early Modern Disputations and Dissertations in an
Interdisciplinary and European Context. Leiden & Boston, MA: Brill, 536–54.
Van Rooy, Raf (forthcoming). ‘A Latin Defence of Early Modern Greek Culture: Alexander
Helladius’s Status praesens (1714) and Its Linguistic Arguments’, in Ioannis Deligiannis
et al. (eds), Post-Byzantine Latinitas: Latin in Post-Byzantine Scholarship (15th–19th
cent.). Turnhout: Brepols.
Van Rooy, Raf and John Considine (2016). ‘Between Homonymy and Polysemy: The
Origins and Career of the English Form Dialect in the Sixteenth Century’, Anglia:
Journal of English Philology 134 (4): 639–67.
Varchi, Benedetto (1570). L’Hercolano: Dialogo . . . nel qual si ragiona generalmente delle
lingue, & in particolare della Toscana, e della Fiorentina. In Vinetia [Venice]: appresso
Filippo Giunti, e Fratelli.
Várvaro, Alberto (2008). ‘La tua loquela ti fa manifesto: Langue et identité dans la
littérature médiévale’, in Moos (2008: 223–36).
Vassallo, Stefano (2005). Himera città greca: Guida alla storia e ai monumenti. Beni
culturali Palermo: Collana di studi e ricerche 8. Palermo: Regione Siciliana.
Vergara, Francisco de (1537). De Græcæ linguæ grammatica libri quinque. Compluti
[Alcalá de Henares]: apud Michaelem de Eguia.
Versteegh, Kees (1986). ‘Latinitas, Hellenismos, ’Arabiyya’, Historiographia Linguistica 13
(2–3): 425–48.
Villani, Francesco Paolo (2003). ‘Scythae: Un problema linguistico, etnografico e culturale
dell’età moderna’, Linguistica 1: 443–91.
Vives, Juan Luis (1531). De disciplinis libri XX. Antuerpiæ [Antwerp]: excudebat . . . Michael
Hillenius.
Vives, Juan Luis (1533). De ratione dicendi libri tres. De consultatione. Louanij [Leuven]: ex
officina Rutgeri Rescij . . . Sumptibus eiusdem ac Bartholomęi Grauij.
Voegelin, Carl F. and Zellig S. Harris (1951). ‘Methods for Determining Intelligibility
among Dialects of Natural Languages’, Proceedings of the American Philosophical
Society 95 (3): 322–9.
Vogt, Joann (1974). The Linguistic Work of Friedrich Karl Fulda. The Hague & Paris:
Mouton.
Vrolijk, Arnoud, Jan Schmidt, and Karin Scheper (2012). Turcksche boucken: . . . The
Oriental Collection of Levinus Warner, Dutch Diplomat in Seventeenth-Century
Istanbul. Eindhoven: Lecturis.
Vuidius, Robertus (1569). ‘De dialectis Græcis libellus’, in Pantaléon Bartelon, De ratione
quantitatis syllabariæ liber . . . Parisiis [Paris]: apud Hieronymum de Marnef, &
Gulielmum Cauellat, 136–48.
Wachter, Johann Georg (1727). Glossarium Germanicum continens origines et antiquitates
linguæ Germanicæ hodiernæ: Specimen ex ampliore farragine decerptum. Lipsiae
[Leipzig]: sumptibus Iacobi Schusteri.
332 
Wackernagel, Jacob (1876). De pathologiæ veterum initiis: Dissertatio

inauguralis . . . Basileæ [Basel]: typis expressit Officina Schweighauseriana.
Waldberg, Max von (1889). ‘Richey, Michael’, in Allgemeine Deutsche Biographie. Munich
& Leipzig: Duncker & Humblot, 28: 436–9.
Walper, Otto (1589). De dialectis Græcæ linguæ præcipuis: Attica; Ionica; Dorica; Æolica; &
coronidis vice nonnulla de proprietate poetica . . . Francofurti ad Moenum [Frankfurt am
Main]: ex officina typographica Ioannis Spiessii.
Walton, Brian (1657). ‘In Biblia polyglotta, prolegomena’, in Brian Walton (ed.), Biblia
sacra polyglotta . . . Londini [London]: imprimebat Thomas Roycroft, 1–102.
Weinreich, Max (1945). ‘ ‫’דער יִיווָא און די ּפרָאבלעמען ֿפון אונדזער ציַיט‬, ‫י ִיווָא בלעטער‬a25 (1): 3–18.
Weinreich, Uriel (1954). ‘Is a Structural Dialectology Possible?’, Word 10 (2–3): 388–400.
Wellisch, Hans H. (1984). Conrad Gessner: A Bio-Bibliography. Zug: IDC.
Wensinck, A. J. (1921). ‘Schultens (Albert)’, in P. C. Molhuysen et al. (eds), Nieuw
Nederlandsch Biografisch Woordenboek. Leiden: A. W. Sijthoff, 5: 707–11.
Whitney, William Dwight (1867). ‘Languages and Dialects’, The North American Review
104 (214): 30–64.
Wichmann, Søren (2016). ‘How to Distinguish Languages and Dialects’, English draft
version of Ran and Wichmann (2018).
Wilkins, John (1668). Essay towards a real character, and a philosophical language. London:
printed for Sa: Gellibrand, and for John Martyn.
Wilkinson, Robert J. (2016). ‘Constructing Syriac in Latin—Establishing the Identity of
Syriac in the West over a Century and a Half (c.1550–c.1700): An Account of
Grammatical and Extra-Linguistic Determinants’, BABELAO 5: 169–283.
Williams, John (1685). A discourse concerning the celebration of divine service in an
unknown tongue. London: printed for Richard Chiswell.
Wolf, Hieronymus (1578). ‘De orthographia Germanica, ac potius Sueuica nostrate’, in
Johannes Rivius, Institutionum grammaticarum . . . libri octo . . . Augustae Vindelicorum
[Augsburg]: Michael Manger excudebat, 594–615.
Wolfram, Walt (1998). ‘Language Ideology and Dialect: Understanding the Oakland
Ebonics Controversy’, Journal of English Linguistics 26 (2): 108–21.
Wood, Robert (1775). An essay on the original genius and writings of Homer, with a
comparative view of the ancient and present state of the Troade. London: printed by
H. Hughs; for T. Payne . . . and P. Elmsly.
Wurm, S. A. and D. C. Laycock (1961). ‘The Question of Language and Dialect in New
Guinea’, Oceania 32 (2): 128–43.
Wyss, Caspar (1650). Dialectologia sacra . . .Tiguri [Zürich]: typis Johannis Jacobi Bodmeri.
Zimmermann, Klaus and Birte Kellermeier-Rehbein, eds (2015). Colonialism and
Missionary Linguistics. Berlin, Munich & Boston, MA: de Gruyter.
Zimont, Elizaveta (2020). ‘Variation diatopique dans la lexicographie français$néerlandais
aux XVIe et XVIIe siècles: Réception et usage des marques diatopiques’, in Van Rooy
(2020b: 48–67).
Zwartjes, Otto (2012). ‘The Historiography of Missionary Linguistics: Present State and
Further Research Opportunities’, Historiographia Linguistica 39 (2–3): 185–242.
Zwinger, Jakob (1605). ‘Græcarum dialectorum hypotyposis . . . ’, in Johannes Scapula,
Lexicon Græco-Latinum nouum . . . Basileæ [Basel]: per Sebastianum Henricpetri,
.1–a.6.
Index
The index records (1) language families, languages, and their varieties, (2) historical figures and
scholars cited in the main text (i.e. not in the footnotes), and in selective fashion (3) terms (in italics),
concepts, subjects, and other entities. I have classified language families, languages, and their varieties
in rather broad categories here in order to avoid having to make countless language/dialect decisions
myself and introducing too many sublevels in the index (two maximum). Only those subdivisions have
been introduced which are relevant to the argumentation of the book, which is why I have, for instance,
refrained from dividing Slavic into East and West.
abolition of dialects 131–2, 288 Aeolic 15, 21, 23, 34, 58–60, 64, 66, 69, 77, 82,
Abstandsprache (‘language by distance’) 271–2 89, 102, 115–16, 141, 143, 175–8, 190,
Académie française 106 222–3
accent 161, 163, 178–9, 193, 195, 197, 279–80 Aetolian 16
accent 268 in Asia Minor 114
vs. dialect 268 Attic 15, 17, 21–3, 59–61, 64, 66, 69–70, 77–8,
accidental (superficial) variation 5–7, 9, 11, 21, 82, 89, 102, 115–16, 123, 128, 141, 143,
33, 59, 74, 76, 80, 104, 109–24, 148, 175–6, 178, 190, 199, 231
157–8, 161, 163–6, 174–6, 179, 195, 197, and High German 206
203, 205, 226–7, 253, 265–6, 282, 298 as the norm of Ancient Greek 222
Adams, James N. 276–7 Biblical 101
ad fontes 28 New Testament 101
Adelung, Johann Christoph 8, 190–2, 196 dialects 4–6, 15–16, 18–21, 23–4, 30–2, 35–6,
Aeolism 34, 178 38, 51–2, 59–60, 63, 65–7, 70–1, 73–4,
Agard, Frederick B. 10, 259–60, 262 77–8, 81–4, 88–90, 98–9, 102, 112,
Ahačič, Kozma 67 115–16, 120, 126, 141–2, 148, 152–3,
Ahrens, Heinrich Ludolf 222 155–6, 165, 174, 176–7, 179, 189–90,
Alans 144 193–5, 199, 205, 207–8, 211–12, 214,
Albertus, Laurentius 101–2, 118 216, 221–4, 296
d’Alembert, Jean le Rond 204 in Byzantine scholarship 197
Alexander the Great 20 classification of 22
Alinei, Mario 2, 47–8, 298 as corrupt 132, 155
Althamer, Andreas 74 dramatic usage of 16
alphabet (letters) 9, 142–3, 174, 178, 186, 188, extinction of 20
195, 197–200, 203 handbooks on 82–4, 102, 107, 125, 179,
American peoples, Scandinavian descent of 145–6 194–5, 203, 206, 208, 211
Amerindian languages 10, 256–7 literary status of 15, 18, 20–1, 23, 130–1,
classification of 281 157, 199, 205, 212, 296
Amerot, Adrien 83–4 modern study of 222
ancestral language 164, 220, 251 and provinces 98–9
ancient Greece 2, 4–5, 296, 299 rediscovery of 81–4
Ancient Greek 4–5, 7, 17, 22, 28–30, 32–8, 43, 49, Doric 15–16, 18–19, 21–3, 59–61, 64, 66, 69,
51, 53–5, 59–60, 62, 65–6, 68–71, 74–5, 77–8, 82, 89, 102, 115–16, 141, 143,
77, 79, 81–4, 88–90, 95, 100–1, 107, 111, 175–6, 190
114–16, 120, 125, 127–8, 130, 132, 143, broadness of 16–17
145, 150, 153–5, 165, 171, 175, 177, 179, Lacedaemonian 16, 66
183, 196, 207, 210–11, 213, 219, 222–3, Rhodian 64
246 and Upper German 206
334 
Ancient Greek (cont.) augment 223

Hellēnismós as ideal 22 Aurelius, Marcus 54
Hellenistic 103, 107 Ausbausprache (‘language by
history of 115 elaboration’) 271–2
Homeric 213–14
Ionic 15–17, 19, 22–3, 59–61, 64, 66, 69–70, Babel, Tower of 40–2, 111, 194, 196, 285, 289
77–8, 82, 89, 102, 115–16, 141, 143, confusion of tongues 194, 199
175–6, 179, 190 Bachmann, Andreas 157
as the ancestral Greek language 214 Bacon, Roger 4–5, 28–44, 47–8, 61, 77, 81, 90, 95,
Chalcidian 16 109, 296
and Low German 206 Baggioni, Daniel 297
Koine 6, 20–2, 33, 59–61, 64, 68–71, 89, 102, Bally, Charles 245
126–9, 141–3, 150, 154, 206 Baltic languages 59, 74, 186
analogical 148 Curonian 74
as normative variety 22, 126 Latvian 60
vs. the other dialects 126, 148 Lithuanian 59–60, 74, 144
poetic dialect 107 Livonian 74
pronunciation 198–9 Old Prussian 59, 74, 144
Byzantine 198 Samogitian 60
Erasmian 199 Yotvingian 60
Renaissance interest in 51–2, 91 Bantu languages 253
as a Scythian dialect 165 barbarismós 20
on Sicily 16 Barbiers, Sjef 265
ancient Greek literature 15, 52, 55, 296 basic vocabulary 8, 186–7, 192, 257, 260, 271
ancient Greek scholarship 199, 223, 296–7, 299 Basque 116, 145, 234–5, 238, 252
ancient Greek tribes 19–20, 60, 76, 101–2, 116 dialects of 238
Hellenistic 103, 107 and Iberian 235
Anderson, Benedict 85 Sara 238
Antesignanus, Petrus 70–1, 91, 95–7, 101–2, Bayly, Anselm 98
154–5 ‘to be’ 187
anthropology 248, 256–8, 294 Beauzée, Nicolas 212
Antikvitetskollegium (College of Antiquity) 159 Bede (the Venerable) 38–9
antiquarianism 183, 199, 221 Belgians 110, 145
Apollonius Dyscolus 21 Bellay, Joachim du 50, 84
appropriation 5, 90–1 Bembo, Pietro 68
approved dialects 70–1 Benzelius, Erik, the Younger 210
Aquinas, Thomas 4, 38–40, 43 Besnier, Pierre 132, 139–40
Aquitanians 110, 145 Bible 28, 30, 39–40, 86, 88, 91, 122, 200, 207,
archaism in dialects 7, 228, 294 220, 239
Aristophanes 16, 20, 54 New Testament 101, 103, 153, 171, 208, 212
Aristotle 6, 18, 54, 57, 109–10, 119, 153 Acts 38, 153, 212
Aristotelian ontology (substance vs. Gothic translation of 162
accidents) 6, 34–5, 43, 109, 123, 148, 282 Mark 171
Aristotelianism and early modern Matthew 39
universities 109 Old Testament 19, 172
Armenian 171 Genesis 40
Armitage, David 3 Judges 39–40, 165
article (part of speech) 34, 74 Proverbs 176
articular pronoun 34 Psalms 38–9
Ascoli, Graziadio Isaia 242, 253 polyglot 122
Asdal, Kristin 3 translation
Ashley, Kathleen M. 90 into German 124, 207
Asia Minor 16, 212–13 into Swedish 207
Auer, Peter 294 Vulgate 30, 38–9
 335
Bibliander, Theodore 58, 77–9, 98, 140, 194 Galatian 24–5

bibliography 53 Gaulish 24, 51, 146
bidialectal speakers 250–1 in Gaul 20–1, 59, 110, 113–14, 144–5
bilingual speakers 250–1 Irish 73
biology 139, 226, 229 dialects of 73
Blair, Ann M. 49, 87 Trierisch 24–5
Blondel, David 164 Welsh 25, 145, 284
Bloomfield, Leonard 248 Celts 110, 145
Blount, Thomas 112, 156 centralized government 69, 96, 99, 106, 128,
Boas, Franz 248 133, 157
Boberg, Charles 279 Ceporin, Jakob 74
Boccaccio, Giovanni 51, 202 Challenge controversy 212
Bodin, Jean 75 Chambers, J. K. 267–8
Boeckh, August 221 Chinese languages, see Sinitic languages
Bolzanio, Urbano 82 Chomsky, Noam 11, 245, 263, 265, 268–70
Bopp, Franz 219–21, 229 Choniates, Michael 20
Borghini, Vincenzo 68 Christina, Queen of Sweden 8, 160, 166
borrowing 285 chronology 183
and conceptual change 80 Chrysoloras, Manuel 81–2
botany 52–3, 139 Chytraeus, Nathan 129
Bourgain, Pascale 37 Cicero 172
Bovelles, Charles de 96 citation practices 300
Bracciolini, Poggio 26 city-states 99
Bréal, Michel 241 civilization 236
Bregius, Johannes 208, 210 Clackson, James 22
Brevini, Franco 201 classical languages 189
bridging context 67 Clement  (Pope) 31
Bringuier, Octavien 241 Clement of Alexandria 19, 55–7, 59, 64–5,
Britons 111 96–7, 108
Bruni, Leonardo 26 Clenardus, Nicolaus 70, 95
de Bry, Johann Theodor 138 Codro 66–7
Buchanan, George 111 cognition 265
bucolic poetry 15 Collegium theologicum 172
Bureau de la statistique 224 Colli, Vincenzo, see Calmeta
Burgundians 144 collocation 5
Burke, Peter 49, 134, 158 Colombat, Bernard 57–8
Byzantine émigrés 81–2 Columbus, Christopher 49
Byzantine scholarship 33, 199, 223, 296–7, 299 Colvin, Stephen 16
common language 5, 50, 59–61, 67–72, 80, 96,
Caesar, Julius 144–5 122, 125–30, 132–3, 150, 164, 204–5
Calepino, Ambrogio 72 as a compilation from different dialects 129
Calmeta 68 ethnic neutrality of 105
Calvet, Louis-Jean 294–5 geographic centrality of 60–1
Camden, William 102 intelligibility of 128
Camerarius, Joachim 78 communicative reach 6
Campbell, Lyle 264 composition (lexical) 285
Canini, Angelo 73–4 conceptual pair, see language/dialect distinction
case 142, 161 conjugation 142, 161
Catholicism 5, 86, 212 Considine, John 85, 144, 162
Caucasian languages 235, 254 consonant 190, 201, 285
Caxton, William 51 cognate 190
Celtic languages 20–1, 24–5, 51, 75, 144, 186, Constantin, Emile 246
190, 234 Constantinople 82
Breton 24 Fall of 52
336 
Cooper, Thomas 72–3 dialect chain 25

Coquebert de Montbret, Charles-Étienne 224–5 dialect chain 276
Coquebert de Montbret, Eugène 224–5 dialect classification 237
core landscape (Kernlandschaft) 242 impossibility of 237
Coseriu, Eugenio 275, 281–2 dialect cluster 260, 262, 283, 301
Cottière, Matthieu 153–4 dialect continuum 231, 250, 257
Council of Europe 290–2 dialect continuum 276
countability of language 5, 88–90, 243, 296, 298 dialect convergence vs. divergence
assumption of 284 252–3, 280
specifically of dialects 89, 296, 298 dialect convergence/divergence 276
creole 234 dialect and corruption 98
Crusius, Johannes 155 dialect and etymology 157
curiosity cabinets 87–8 dialect as a fictitious concept 239, 246
customs 179 dialect geography 9, 228, 230, 243, 298, 300
Cysouw, Michael 277, 282 dialect as an identity mark 70
dialect knowledge, utility of 207–8
Dacus, Johannes 34 dialect levelling 276
Dahan, Gilbert 42 dialect loss 276
Dante Alighieri 25–7, 42, 99, 202 dialect and migration 138
Darwin, Charles 226 dialect mixture 16, 51, 241
De Laet, Johannes 145–6 dialect as a popular concept 238
declension 161 dialect preservation 210
deep structure 259, 263–5 dialect as a relative concept 223, 235, 238,
degenerare in dialectum 165 246, 279
degenerate offshoot 8, 176–8, 182, 211–12 dialect renaissance 281
degeneration of language 167 dialect shift 276
degradation of dialects 289 dialect and social class 130
Delbrück, Berthold 229 dialect, sociological view of 248, 274
derivation 190 dialect status, acceptance of 271
Descartes, René 204 dialect term/word 276–7
dia- 275 dialect and writing 207
diacritic marks 193 dialectal 275
dialect 4, 11, 19–20, 58, 73, 75, 90, 98, 112, dialecte 129, 131, 151–2, 212, 246–8
153–4, 188, 220, 227, 249–52, 254–5, naturalization as a French word 151
266–8, 270, 275–85, 297, 299–300 dialectologia 9, 208, 210
abandonment of the term 251, 275, 299 dialectological tradition 9, 208, 211, 216
country 133 dialectological turn 10
compositions with 275–7, 282 Dialectologie 210
daughter 136, 139, 247 dialectology (institutionalized) 210–11, 225, 228,
definitions of 11, 251 230, 245, 263, 266–7, 273, 279, 281,
derivations of 11, 275–8, 282 293–4, 298
as a difficult word 285 continental European 279
in lay usage 283 dialectology 21, 208, 275
naturalization as an English word 73 externalist form of 249
as a non-technical term 249–51 Gallo-Romance 241, 243, 247, 251
poorly defined 251 generative 245, 264
as a term for Greek only 154 German 242
dialect area 267 Italian 242
dialect area 276 Montpellier school of 241
dialect attitudes 17, 61 Parisian school of 242
dialect boundaries 10, 239, 254, 300 structuralist 245, 247–55, 265, 298
arbitrariness of 10, 300 Swiss 249
problematic nature of 25, 245, 250, 267, traditional 263, 267, 281
276–7, 300 dialectometry 276
 337
dialects, existence of diglossia 278

lack of 251 Diogenes Laertius 17
‘only dialects exist’ 247 Diogenes of Babylon 17–20, 98
dialectus 4–5, 7–8, 24, 35, 53, 55–81, 84, 90, diversitie 73
95–8, 101–3, 107–8, 120–3, 130–1, Driessen, Anton 173, 175
137–42, 153, 156, 167, 180–1, 194, 204 Drusius, Johannes 103
communis 60, 89, 150 dynamism of language 239
egkhṓrios 116
ethnikḗ 116 École des Hautes Études 247
generalis 115 edit distance, see Levenshtein distance
naturalization as a Latin word 53, 62–4, 149 (normalized)
polysemy of 57–8, 80, 103, 121 Egyptian 20, 164, 220, 253
primaria 116 Coptic 253
princeps 115 Eichhorn, Johann Gottfried 190
principalis 116 Eisenstein, Elizabeth L. 52
propinquior 116 Eliot, Thomas 72–3
remotior 116 Eloy, Jean-Michel 278, 282
secundaria 116 encyclopaedia 204
specialis 115 Encyclopédie 204, 212
topikḗ 116 ending 142, 161
urbica 116 English 2–4, 25, 33, 36, 38, 51, 69, 72–3, 75, 106,
vulgaris 60 133, 156, 175, 184, 214, 216, 227, 265,
dialégomai 18 285, 292–3
Dialekt (Dialect) 216, 220, 222, 225–6, American 266, 285–6, 289
235–6, 286 African American Vernacular 267, 290
Hauptdialekt 116, 236 on Martha’s Vineyard 266
Nebendialekt 116 in New York 266
Unterdialekt 236 Pacific Northwest 266
diálektos 2, 4–5, 17–21, 23–4, 27, 43, 55–7, Pittsburgh 266
63, 65, 72, 76–7, 80–1, 90–1, 96, 98, British 285–6
101–3, 108, 147, 149–50, 154–5, Eastern 89
208, 297 London 133
ethnic conception of 17, 19, 27, 297 Northern 89
Latinization of 90, 149 Southern 89
literary conception of 20, 27 Western 89
and lower class speech 20 common language 89, 106, 127–8
appropriation of 149–50 of the court and intelligentsia 128
parátonos 20 dialects of 89, 127–8, 156, 206, 290
as ‘particularity of tongue’ 18 diphthong change in 266
polysemy of 153–4 lexicon 271
spatial conception of 17, 19–20, 27, 297 Middle English 31
dialetto 202, 280 Devonish 72
dialettuzzolo 116 East Midland 31
diaphasic 275 Kentish 72
diastratic 275 Northern 32, 72
diasystem 10, 247–51, 254, 259, 272 South Eastern 51
diatopic 275 South Western 31
Dictionnaire de Trévoux 197 Southern 72
dictionary 185 West Midland 31
dialect 199 Poetic 89
Diderot, Denis 204 received pronunciation 280
‘to differ only in dialect’ 5, 74–6, 144–6, 164, 166, as a Romance language 177
220, 226, 290, 292 standard language 280, 290
expansions of the phrase 109–13 variability of 51
digamma 34 Welsh 284
338 
Enlightenment 8, 192, 201, 203–4, 206, 216, social variation in 130

286, 297 standard language 88, 252
Ephraimites 39–40, 165, 239 standardization of 96
Erasmus, Desiderius 65–6 Walloon 35–7, 117
Eskhult, Josef 173 Occitan (langues d’oc) 26, 241
Estienne, Henri 98, 130, 195–6 Gascon 117
ethnography 183–4, 191–2 Limousin 89
of Asia 220 Provençal 36, 89, 117, 161
Ethnologue 10, 258, 261, 277 Toulousian 161
éthnos 19, 101 Gatterer, Johann Christoph 8, 183–93, 196, 200,
etymology 177, 189, 296 203–4, 286, 297
European Charter for Regional or Minority Gauchat, Louis 242
Languages 278, 290–2 Gaza, Theodore 82
European Union 292 Gedike, Friedrich 206–8
exegesis 4 Geeraerts, Dirk 276
explorations, age of 49–50 Geijer, Erik Gustaf 160–1
extinct languages 125 gender studies 276
generic vs. specific dialects 205
feudalism 289 genius 180, 207
Filelfo, Francesco 82 gens 101
Flavio, Biondo 26 geography 37, 183, 238–9, 275
Florence 81 ethno- 147
Franciscans 30 geologia 208
Franks 144 Gerald of Wales 25
French National Convention 288 Gerard of Huy 30
French Revolution 225, 288–9 gergo 102
Fulda, Friedrich Carl 9, 214–16, 224, 230 German 32–3, 47, 60–1, 68–9, 73–4, 79, 91, 97,
Furetière, Antoine 129, 131, 151–2, 155 104, 115, 118, 122, 129, 150, 153, 165,
Fust, Johann 91 183–4, 190–1, 200, 205, 207, 210, 214–16,
220–1, 227, 233–4, 237, 267, 272
Gabelentz, Georg von der 229–30, 286–8, 299 Austrian 186
Galiani, Ferdinando 9, 201–3 Bavarian 61, 68–9, 186, 286
Gallo-Romance 33, 35–6, 39–40, 96, 278 dialects 221
dialects 241 and Bible translation 207
Franco-Provençal (Arpitan) 242, 278 Büchersprache 185
Lyonnais 89 common language 105, 124
Savoyard 196 dialects of 78, 89, 101–2, 115, 124, 155, 161,
French (langues d’oïl) 4, 26, 33, 35–8, 69–71, 198, 206, 212, 214–15
79, 84, 88–9, 96, 114–15, 117–18, 120, as conservative 185
128–30, 139, 151–2, 154, 165, 175, 177, in parallel to Greek 206
195–6, 214, 225, 241, 252–3, 269, 288–9 and social class 155
Burgundian 32, 35, 37 Franconian 115, 153, 186
common language 70 Hamburg 210
and Corsican 252, 271 High 102, 118, 123, 132–3, 150, 206–7,
dialects of 70–1, 98, 117–18, 130, 161, 195, 210, 287
211, 269 and Attic Greek 206
of the Île-de-France 51 at the imperial court 105
lingua Gallicana 33 Low 102, 118, 129, 132, 153, 207
Norman 35, 98, 161 and Ionic Greek 206
of Orléans 115 of Meissen 115, 153
Parisian 37, 40, 70, 89, 128, 130, 161 Osnabrück 210
patois 212 Palatine 286
Picard 35–7, 40, 89, 98, 161, 278 Plattdeutsch 286–7
at the royal court 70, 89 of the Saxon chancellery 69, 124
 339
Schriftdeutsch 286 glottopolitics 271, 293

Silesian 153 Goebl, Hans 241–2, 276
in Speyer 105 Goldhagen, Hermann 98–9
standard language 69, 261 Good, Jeff 277, 282
Swabian 68–9, 120, 153, 186, 286 good language 98
Swiss 47, 61, 73, 157, 261 Gooskens, Charlotte 258–9
Thuringian 115, 153 Goropius Becanus, Johannes 6, 119–22, 257
Ubian 68–9 Göschel, Joachim 281
Upper, and Doric Greek 206 Goths 144
Germanic languages 25, 42, 49, 60, 75, 100, 113, Göttingen Historical Institute 183
116, 137, 144–5, 149, 157, 186, 190, Gousset, Jacques 173, 175
215–16, 225 Grafton, Anthony 91
Danish 214, 221, 261, 267, 292 grammar 11, 34, 100, 147, 151, 259–60, 263,
as a dialect chain 25 267–8, 271, 279, 301
Dutch 69, 100, 111, 119, 121, 148, 157, 201, Byzantine 30, 126
212, 216, 280 comparative 185
Brabantian 61, 120 Greek 34, 76, 82
common written 148 of individual dialects 264
dialects of 148 Latin 34
Flemish 120 manuals 267
Guelderish 157 modistic (speculative) 25, 34
interlanguage (tussentaal) 280 panlectal (polylectal) 267
kinship with Persian 201 universal (generative) 263, 265
as the primeval language 119, 121 as a universal art 34–5
English, see English Greek, see Ancient Greek, vernacular Greek
German, see German (early modern), Modern Greek
Gothic 162, 165 history of 152
as the Germanic mother language 214 Greek Orthodox Church 123
grammar 215 Greek studies 69, 71, 91–2, 126, 207, 211
lexical roots 215 Greene, Robert Lane 292–3, 299
Norwegian 267, 292 Grégoire, Henri (Abbé) 288–9
Old Saxon 165 Gregorius Pardus, see Gregory of Corinth
protolanguage 200 Gregory of Corinth 19, 72, 82
in Scandinavia 267 Grenzsaum 272
Swedish 121, 160, 162, 165, 207, 227, 261, 267 Gretser, Jakob 194
and Bible translation 207 Grimm, Jacob 220
Dalecarlian 164 Groddeck, Benjamin 181, 206
dialects of 161, 183, 210 Grotius, Hugo 145
norm 207 Gubbio tablets 51
orthography 207 Guillén de Brocar, Arnao 101
as a Scythian dialect 162 Guldi, Jo 3
Västergötland dialect of 121, 207 gunpowder 28
Gessner, Conrad 4–5, 47–62, 64–5, 72–5, 77–9, Gutenberg, Johann 91
88, 91, 96–8, 113, 137, 140, 144, 296
Giard, Luce 146 Habsburg 100
Giese, Albert 9, 221–4, 230 Halle, Morris 264
Gileadites 39–40, 239 Haloinus, Georgius 99–100, 113–14
Gill, Alexander 89–90, 133 Ham 79
Gilliéron, Jules 10, 241, 251, 275 Hamitic languages 253
Gipper, Sonja 67 Handbook of dialectology 278–80
Glarean, Heinrich 59 Harris, Zellig 10, 256–7, 262–3
glôssa/glôtta 19 Haspelmath, Martin 277
glossonym 277, 282 Haugen, Einar 281, 284
Glottolog 10, 258, 277–8 Heinsius, Daniel 103
340 
Helladius, Alexander 122–3 imperialism, ideology of 294

Helm, Abraham 125, 132–3, 204–6, 211 Indo-European languages 219–21, 225, 228,
Hemsterhuis, Tiberius 180 230–1, 237, 245, 252, 298
Henry, Victor 240–1 as dialects 248, 250
Heraclitus 239 inductive method 28, 31
heraldry 183 info-lust 87–8
Herodotus 15–17, 19, 78 information explosion 4–5, 84, 87–8,
Hervet, Gentien 55–7 91, 296
Hesychius 20 innateness of language 263
Heylyn, Peter 156 intellectual history 2–3
Hickey, Raymond 279 Internet 11, 293
Himera (Sicily) 16 isogloss 241, 249–51
Hinskens, Frans 279 Italic languages 223
Hippocrates 54, 70 Latin, see Latin
Hirschi, Caspar 85 Umbrian 51, 226
historiography 8, 15, 147, 151, 159, 183–92, 204, Italo-Romance 26, 99, 202, 280
214, 220, 294 Bolognese 99
and statistics 189 variation in the city of Bologna 99
universal 183 Corsican 251–2, 271
history, sense of 146 Italian 26, 33, 68–9, 79, 81, 91, 115, 120, 128,
Hockett, Charles F. 254–6, 258, 266 139, 165, 175, 251–2, 269, 271
Hof, Sven 121, 207–8 common 202
Hoius, Andreas 97, 154–5 as corrupted Latin 174
Homer 15, 21, 54, 153 dialects of 81, 161, 202, 212
language of 213–14 Florentine 115
Horace 23 standard language 280
van der Horst, Joop 88 Tuscan 128, 161, 202, 251, 280
Hottinger, Johann Heinrich 7, 141–3, Lombard 161
213, 249 Neapolitan 161, 201–2
Hovdhaugen, Even 43 Roman 238
Howell, James 114, 116–17, 138–9 Sicilian 280
Howes, Thomas 214 Venetian 161, 238, 280
Howlet, Richard 73
humanism 2, 5–6, 26, 87–8, 100, 296–7, 299 Jablonski, Paul Ernst 212–13
Hyde, Thomas 198 Jäger, Andreas 166
Japheth 79
Iberian 235 jargon 102, 254, 278
and Basque 235 jargon 132, 212, 275
Ibero-Romance 97 Jaucourt, Louis de 97
Catalan 261 Jerome 24–5, 30, 39, 49, 98
as a Spanish dialect 252 Jesus Christ 49
Spanish 32–3, 51, 69, 79, 115, 120, 128, 139, Jiménez de Rada, Rodrigo 41–2
175, 261 John the Grammarian (also known as John
Castilian 51, 128 Philoponus) 19–22, 33, 77, 89, 193–5,
of Toledo 115 199, 203
ideal speaker 245, 265, 280 Jones, William 219–21
idiolect 226, 248, 255–6, 275, 277 Joseph, John E. 85
idiom 5, 35–6, 42–3, 76–8, 179, 194, 197
idiom 227 Kals, Jan Willem 116, 180–2
idioma 4, 31–8, 42–3, 76–7, 90, 131 ten Kate, Lambert 116, 148–9, 151
idíōma 77 Keller, Raffael 53–4
idiome 212 kharak-root 16–17, 19
idiota 38 kharaktḗr 16
Ihre, Johan 183 kharássō 17
 341
Kirchmaier, Georg Caspar 155, 166 as a flexible matrix 150–1, 228, 242, 254–5
Kirchmaier, Georg Wilhelm 206 future of 299–302
Kloss, Heinz 271–3, 281–3 historical status of 12
koine 71 implicitness of 248
koinḗ 60, 84, 89 interpretations of 92, 147–51, 156, 167, 202,
as a language/dialect criterion 252 211–12, 214, 224, 228, 242, 248, 280, 282,
Kretzschmar, William A., Jr 273–4, 276, 281–2, 289, 297, 299
300–1 language-historical (genealogical) conception
Kristiansen, Gitte 276 of 7–9, 58, 61, 78–9, 136–46, 148–9,
Kristiansen, Tore 279 151–2, 164–5, 167, 176–7, 182, 189, 205,
Kümin, Beat 99 213–14, 216, 219–22, 225–7, 229–30,
Kunstsprache 15 247–9, 251, 298
Kurath, Hans 242, 267 mother/offshoot distinction 213
lay conceptions of 11, 186, 252, 254, 256, 259,
Labov, William 266–7, 279, 281, 290, 293 266, 272, 276, 283–95
language 11, 19, 73, 127–8, 130, 220, 227, 252, and linguistic features 163, 247, 256–7, 259,
254–5, 266–70, 277–8, 282–4, 268, 271, 276
297, 300 non-scientific nature of 268
common 127 normative conception of (analogy vs.
in lay usage 283 anomaly) 6–8, 125–7, 148, 176, 247, 251,
mother 136, 138, 150, 152, 196, 247 279, 286, 294
as a non-technical term 268–9 as an observational artifact 273
poorly defined 251 political activations of 11, 287–93
radical 112, 156 politicization of 11, 105–6, 158, 288–9
language/dialect distinction problematic nature of 291
absence of 17 and preconceived ideas 147
alternatives for 270 quantification of 258, 260–2
ambiguity of 215, 229, 250, 279, 284, 287, 294 relativity of 247–8, 252
ambiguous attitudes towards 248, 282–3 as a remnant of earlier thought 273, 282
arbitrariness of 9–10, 182, 189, 226, 228, 231, as the responsibility of linguists 252, 301–2
248, 298, 300 scepticism about 212–16, 224–5, 228, 230,
Aristotelian conception of 6–9, 33–4, 109–24, 240–2, 247, 249, 282, 284, 297–8, 300
147–9, 156–7, 163–4, 166–7, 174–6, 182, scientific redefinition of 251–4, 260, 282–3,
195, 197, 205, 215–16, 227, 230, 253, 298, 300
264, 282 as a self-evident given 1, 12, 87–8, 92, 113,
attributed to the Greeks 120 123, 149, 166, 203, 220, 229, 242, 298–9
auxiliary status of 162, 182, 230, 296 and social class 60, 134
and communicative reach 122–3 (socio)political conception of 7, 11, 157–8,
consonants vs. vowels 201, 203 270, 274, 282–3, 289–90, 293–5, 298, 300
convenience of 287 rejection of 287, 292, 294
and cut-off point 253, 260–1, 271, 283, 300 spatial conception of 5–9, 33, 59, 95–101,
debate about (or lack thereof ) 249, 280–1, 147–9, 151, 156, 165, 176, 228, 230,
298–9 266–7, 276, 279, 282, 298
de-Hellenization of 7, 133–4, 151–7, 167, 176, subjective conception of (status) 6, 125–35,
190, 194, 203, 211, 297, 299 148–9, 151–2, 157, 165, 195, 202, 225,
diagnostic criteria for 6, 107, 123, 147, 249, 279, 288–90, 293, 298, 300
165, 167 subsumption of dialect under language 18,
as a discursive strategy 7, 143–6 47–8, 57, 59, 62–4, 67, 69, 72–3, 76, 80,
dismissal of 252, 268–70, 273–4, 278–80, 92, 96, 101–3, 107–9, 120, 128, 147,
282–3, 295 154–5, 205, 223, 297
emergence of 85, 90–1 supplementing of 11, 211–12, 248, 252, 254–5,
uncoordinated nature of 90 270–3, 280, 282
ethnic conception of 6, 33, 59, 101–8, 147–9, and syntax 202–3
151, 156, 205, 224 tenability of 211
342 
language/dialect distinction (cont.) language system 10–11, 245, 247, 253–4, 265
as a useless abstraction 10, 216, 240, 270, 274, as closed 265, 273–4, 276, 279, 301
293, 300 as complex and self-organizing 273, 282,
value of 215, 229, 246, 258, 270 300–1
language/dialect status homogeneity of 10, 250, 280
determination of 10, 195, 273, 277–8, 287 horizontal vs. vertical varieties 253–4
forced 291 idealized 269, 281
by laypeople (subjective perception) 256–7 mixture of 269
debates about 290–3, 298 subsystem 253–4
preconceived ideas about 298 regional 253–4
language atlas 239, 298 language variation 111, 116, 119, 139, 149, 232,
language attitudes 43 235, 241, 247, 249, 252, 255, 263, 265–6,
research 293 270, 274–5, 278, 280–1, 285, 289–90, 296,
language catalogue 4, 10, 47, 52–3, 75, 88, 113, 298, 300–1
258, 277–8, 296, 298 awareness of 16–17, 27, 64
language change 7, 111, 195, 197, 232, 235, and boundaries 88, 231, 236–9, 254, 281, 300
237, 252 arbitrariness of 238–9, 300
causes of 164, 205–6, 285, 301 fuzziness of 246, 272, 300
colonization 17, 201 categorization of 88, 289
isolation 286 need for 270
language-external 221 in one city 99, 266–7
migration 201 classes of 22, 178–80, 182, 193–203, 206,
space 164, 201 268, 297
time 164, 201, 205, 215 as a continuum 88, 114, 239, 254
constant nature of 245 curiosity about 51, 79, 183
as a cyclical process 7, 161, 164, 200–1, 227, different levels of 115, 122–3, 205
249, 252 discovery of 49
and ethno-geographic dispersion 20 evaluation of 61
gradual nature of 146, 161, 176, 181, 200–1, and external factors 11, 212, 247–8, 252–3,
205, 222, 226–7, 301 258–9, 270–2, 275–6, 279, 287, 289, 298
tree model (Stammbaumtheorie) 10, culture 247–9, 254, 258–9, 267, 272
231–3, 236 ethnolinguistic identity 258, 267
wave model (Wellentheorie) 10, 231–3, exclusion of 253
236, 247 geography 248–9, 267, 276
language classification 4–5, 7, 139, 221, 233, history 267
281–2, 284, 296, 298 perception 276
language cluster 260, 262, 283, 301 politics 247, 252, 254, 259, 267
language community 287 school use 272
language comparison and social class 248–9, 252, 259, 267, 276
historiography 183–92 style 276
language contact 236–7, 239, 249, 253, 301 time 249
language and diplomacy 131 gradational conception of 6, 99, 114–17, 123,
language and ethnic history 191–2 148, 187–8, 205, 229, 235–6, 246–7, 252
language family 229, 231, 241, 277 in the Holy Roman Empire 122
language history 7–8, 79, 122, 136–46, 149, 162, lack of interest in 24–5
167, 215–16 in the New World 145–6
language as living organism 226, 242 New Spain 145
language mixture 234, 237, 239 origin of 200
language origin 162, 166, 194–5 universality of 5–6, 62, 99–101, 161, 227
language and politics 269 as unrestrained 197
language questions 127 langue 10, 245–7
questione della lingua 51, 68 languoid 11, 277, 282
language as a social fact 266, 271, 275 Larramendi, Manuel de 116
language and state 105–6, 212 Lascaris, Constantine 82
 343
lāšṓn 103 lexicostatistics 8, 10, 183–92, 257, 283, 286,

Latin 2, 4–5, 15, 20, 22–6, 28, 30, 32–6, 38, 47, 297, 300
49, 51, 53–4, 62–4, 66, 68, 72–3, 79–81, cut-off point 10, 189, 283
84–5, 100–2, 111, 113–14, 129, 139, léxis 17–18
145, 171–2, 174–8, 190, 198, 208, lingua 4–5, 7, 31–9, 42–3, 57–8, 62–81, 90, 95,
210, 213, 219, 222–3, 233, 246, 102–3, 120–1, 131, 167, 181, 194
276–7 communis 5, 60, 67–8, 80, 84
as a daughter language of Greek 213 as language of learned men 60
as a dialect of Greek 177–8, 223 communissima 69
dialects of 175 diversa 202
as a lingua franca 63, 80, 84, 145 ipsissima 150
literature 296 polysemy of 62
as a living language 33 primigenia 113
Neo- 5, 63, 74–5, 85, 149 principalis 61
orthographies of 32–3 tota 111
Paduan 100 linguagium 60
periodization of 66 linguistic corruption 177, 212
Praenestine 23 linguistic distance (Abstand) 271, 273,
pronunciations of 32–3 282, 301
regional variation in 23–4, 38, 206, 276–7 bimodal distribution of 10, 260–2, 283
registers in 23, 78 as a criterion 200, 258, 271
social variation in 24, 100 quantification of 8, 10, 189, 257, 259–61,
as a uniform language 62, 66, 100 281, 283
and vernacular 26 linguistic horizon 43, 49–52, 183, 296
Vulgar 233, 237 linguistic kinship 191, 195, 220
vowel system 233, 235–6 degree of 187–90, 301
lect 11, 275–6 as a continuum 189, 193
communa- 275 as a language/dialect criterion 251–3
docu- 11, 277, 282 quantification of 256–7, 276, 301
ethno- 275–6 linguistics 2, 11, 240, 244, 272, 276–8, 300–1
multi- 275 as an autonomous field of study 2, 9, 225,
gender- 275–6 229–30, 266, 293, 298–9
idio- 275 basic assumptions of 242
natio- 276 cognitive 276
as a neutral term 275 computational 282
regio- 275 contact 234, 249
socio- 275 creole 234
techno- 275 descriptive 259
lectometry 276, 281 dialectological turn in 263
Leibniz, Gottfried Wilhelm 185 general 266
Leskien, August 222, 232 generative 11, 244–5, 259, 263–8, 280–2, 298
letter mutations (permutatio litterarum) 21–2, Minimalist programme 265
58–9, 92, 142–3, 161, 163, 165–6, Optimality Theory 265
178–9, 187, 189, 191, 193–4, 196–7, Principles and Parameters theory 265
199–200, 205 geographical 246
and etymology 197 historical-comparative 9, 139, 219–23, 225,
Leu, Urs B. 53–4 228, 230, 245, 248, 252, 298
Leusden, Johannes 150 history of xiii, 80, 225, 278, 299–300
Levenshtein distance (normalized) 261 vs. lay ideas 277, 280, 293–4
lexeme 276 linguistics 211
lexicon 7, 9–10, 27, 111, 142, 161, 163, 165–6, mainstream 243, 245, 274, 280
172, 179, 189, 193–200, 202–3, 221, missionary 49–50
253–4, 257, 259–61, 268, 279, 286, 301 modern 186–7, 219, 268, 299
as language/dialect criterion 271 and other disciplines 294–5
344 
linguistics (cont.) Mayer, Bartholomaeus 199–200

pre-comparative 143–4, 146, 177, 186, 188 Mazzarella-Farao, Francesco 116
social turn in 253, 281 Meillet, Antoine 248, 252–3
socio- 11, 186, 244, 259, 263, 265–70, 272–3, Melanchthon, Philipp 68–9, 74, 80, 91
276, 279, 281 Merton, Robert K. 109–10, 155–6
externalist approach of 268, 281 Merula, Paulus 110–11, 113, 136–8, 140,
structuralist 10, 244–60, 263, 266, 280–1, 298 144–5
in the USA 255, 275 metalanguage 16, 18, 24, 31, 35, 80, 131, 136,
textbooks 248, 281 139, 147, 149–50, 167, 173, 238, 262, 278,
typology 277 283, 289, 299–300
variationist 293 metaphor 178
Lipsius, Justus 201 family 139, 141, 215–16, 226, 298
literary dialects 250 Meyer, Georg 142
literary language 254 Meyer, Paul 241–2, 245–6, 274
littera 198 Miechowita, Maciej 60
Livy 100 Milton, John 89
loan translation 60 minority languages 271, 278, 290–2
local dialect 248 Mischsprache 223
locutio 39 Mithridates , king of Pontus 52, 54, 113–14
lógos 208 Modern Greek 263, 280
Longobards 144 dialects of 280
longue durée 3 Cappadocian 280
loquela 38–9 Cretan 280
Lord’s Prayer 157 Cypriot 280
Lucian 66 phonology 263
Ludolf, Hiob 157–8 regional accents 280
Ludovici, Michael 157 standard 280
Luther, Martin 69, 71, 86, 96–7, 118, 124 Mokyr, Joel 204
Lycaonian 212–13 monogenesis of language 78–9, 187–8, 200
as an Ancient Greek dialect 212 Morhof, Daniel Georg 132
lyric poetry 15 morphology 27, 142–3, 163, 190–1, 195, 197–8,
200, 202, 221, 278, 301
Maaler, Josua 47 inflection 199–200
Macaulay, Ronald K. S. 266–7 as language/dialect criterion 198
Machan, Tim William 85 morphosyntax (language/dialect
macro- vs. micro-parametric variation 265 criterion) 10, 253–4
Malay language 261 Morpurgo Davies, Anna 228, 231–3, 240, 242
Indonesian 261 Morrhy, Gérard 83
Malaysian 261 Mosellanus, Petrus 72
Malherbe, François de 127 Moser, Ingunn 3
Manutius, Aldus 99 Müller, Andreas 157
de la Mare, William 30 Müller, Max 227
Martial 62 multidialectal outlook 51, 84, 296
Martinet, André 10, 249, 251–3, 282–3, 301 multi-linguis 113
Martini, Johann Georg 117, 206 Mundart 220, 222, 225–6, 235, 238, 286
Mass, language of 122, 127, 212 Hauptmundart 116
mass education 11, 288–9, 293, 301 Untermundart 235–6
mass media 288–9, 293 musicality of language 239
Matthias de Miechow, see Miechowita, Maciej mutual intelligibility 6–7, 9–10, 60, 117–23, 158,
Max Planck Institute for Evolutionary 164, 166–7, 176, 201, 215, 256–62, 271–2,
Anthropology 277 282, 286, 301
Maxwell, Alexander 268, 270 asymmetric nature of 258, 292
May, Stephen 289 as a contextual factor 261
 345
as a language/dialect criterion 10, 60, 118, number of languages (total) 55–6, 64–5, 88, 229,
147–8, 161, 229–30, 246, 248–9, 255–62, 261, 272, 284
267, 271, 287, 292, 294 infinity of 221
arbitrariness of 229, 235–6, 246, 256, 287 numeral 187
in laypeople’s ideas 256, 287, 299
objectivity of 292, 299 Oakland Ebonics controversy 11, 290, 293
rejection of 236, 257, 261, 271 obliteration by incorporation 109–10, 155–6
utility of 246, 259 Ogilvie, Brian W. 47, 53, 296
vagueness of 215 optics 28
and good will 258–9 Oriental studies 173
gradational nature of 235, 246, 258–9, 292 orthography 202, 207, 267
immediate 6, 119–22, 124 Otfrid 185–6
lack of 6, 16, 41, 75, 96–7, 118, 164, 269, Ottomans 171
286–7, 301
in relation to social class 118 pánta rheî 239
measuring of 271 paradiastrophḗ 153
objective definition of 229, 287 paragōgḗ 16
and previous knowledge 258 Paris, Gaston 241–2, 245, 274
as strictly binary 122, 124, 164 parole 10, 245, 247
as a surface phenomenon 259 pars grammaticalis 198
testing 10, 256–9, 282 parts of speech 74
threshold 259 Pascale, Carlo 131
and uneducated speakers 287 Pasor, Georg 101, 208
mutual laughter 36, 43 Patavinitas 100
mutual recognizability 272 pathology 21, 59, 197
Mylius, Abraham 7, 111–12, 121, 140 patois
Myricaeus, Johann Caspar 121 annihilation of 288–9
mythology 185 as corrupt 98, 128, 131, 212
patois 97–8, 132, 224–5, 227, 241, 251–2
Nahuatl (Mexican) 145 negative connotations of 252
Napoleon 225 as pure 132
nation 6, 33, 43, 101–8, 241, 285 as remnants of feudal society 289
and dialect 101–5 as supplement to the conceptual pair 252, 282
natio 33, 43, 101 patriotism 69
nation 212 and dialect description 210
nation-state 85, 267, 276, 288–9, 292–4, 301 Paul, Hermann 229, 253
national community 84–5 Pelasgian (as the ancestral language of
national language 85, 252, 289 Greek) 214
nationalism 85, 287–9, 292, 296 Pennisi, Antonino 202
native dialect 70 Perotti, Niccolò 62–3, 100
natural history 227 Persian (Indo-Iranian language) 49, 227
Renaissance invention of 296 kinship with Dutch 201
natural language 247 Persons, Robert 196
development 228, 273 Peter (apostle) 39
natural philosophy 31 Peters, Manfred 57–8
nature 239 Petrarch 51, 202
neogrammarians (Junggrammatiker) 9, 228–9, Peucer, Caspar 144
234–5, 241 peuple 212
New Guinea, languages of 257 Pfeiffer, August 117, 206
Newton, Brian 263–4 Phillips, Edward 106
Nibbe, Johann Barthold 210 philology 2, 8–9, 11, 20–2, 27, 147, 151, 155, 159,
Nicholas of Lyra 40 183, 194–5, 199, 204, 206–7, 211, 215,
normative linguistic thinking 22, 128 219, 221, 233–4, 237, 245, 284, 289, 301
NORMs 228 phōnē ̀ eggrámmatos 18, 20
346 
phonetics 27, 202, 238–9, 253, 257, 268, 279 prosody 179
phonology 10, 259–60, 268, 271, 279, 301 Protestantism 5, 85–6, 92, 207, 212
and structuralism 249 Proto-Indo-European 163, 221, 225, 231
phraseology 142–3, 161, 163, 174, 178–9, 195, 202 as unitary 233
Phrygian 213 vowel system 245
as an Ancient Greek dialect 213 protolanguage 173, 221, 280
physiology 265 reconstructed 225
Piccolomini, Enea Silvio (Pope Pius ) 25 prototypicality 276
Pillot, Jean 130 proverbs 196
Pindar 22 province (provincia) 98–9, 165, 167, 212, 215,
Plato 16, 57, 153 241, 247
Plautus 23 provincial dialect 252
Plesch, Véronique 90 purity of dialects 294
Pliny the Elder 75 Puttenham, George 130
Plutarch, pseudo- 21
poetical license 285–6 quadrivium 28
Polák, Václav 10, 251, 253–4, 257, 264 Quechua 264
Polier de Bottens, Antoine-Noé de 181 questione della lingua, see language questions
Pollux, Julius 17 questionnaire 245, 293
Polybius 75 Quintilian 23–4, 59, 63, 72, 78, 90, 100
polyglot 51, 113–14, 123, 143
Poole, Matthew 100 race 241
Pop, Sever 251, 253 Raphelengius, Franciscus 201
popular knowledge 2 Rask, Rasmus 219–21
popularization of knowledge 285, 290, 293 Rastell, John 75–6
popolo 102 rationalism 8, 192, 203–4, 208, 211, 216, 286, 297
populus 101 Ravis, Christian 8, 112–13, 141, 166, 195–6, 199
Porphyry 54 Ray, John 133
postcolonial studies 276, 294–5 Reconquista 51
Pott, August Friedrich 222, 229, 240 Reduction 187–8
Prague linguistic circle 253 Reformation 85–6
prehistory 8, 185 regional dialects 280
prescriptivism 290, 294 register 275
primary dialect 277 Reland, Adriaan 171
vs. secondary 116 related dialects 6, 188–9, 202–3, 205, 222, 224,
primeval language 177, 181, 200 235, 248, 253, 256, 264, 271, 283, 286, 300
principle 196 related languages 187–90, 222, 224
print-capitalism 85 Renaissance 2–4, 6, 27, 51–2, 62–4, 71, 80, 84–5,
printing press 5, 51–2 87, 91, 106, 126, 146, 156, 167, 193–4,
commercialization of 5, 51 230, 282–3, 289, 296, 298, 301
invention of 90–1, 296 collector’s mania 87
and standardization 85–6 conception of space 99
Priscian 34 Italy 21, 24, 26
Prodigal Son, Parable of 225 multilingual perspective of 52
pronoun 187 scholarship 297, 301
demonstrative 34 Renan, Ernest 227
Greek 82 Republic of Letters 51, 63, 90–1
pronunciation 7, 9, 33, 111, 157, 161, 163, 166, Rhellicanus, Johannes 58, 75, 79
174, 178–9, 187, 194–200, 202–3, 216, rhetoric 15
221, 241, 268, 285 Rhodes 15, 64
common 161 rhythm of language 239
correct vs. incorrect 241 Ribier, Guillaume 122
vulgar 157 Rice, Eugene F. 91
propago 61, 137, 152, 196, 213 Richey, Michael 210
 347
richness of dialects 130, 208, 210 Schuchardt, Hugo 10, 231–43, 249, 253, 274, 298
Rider, John 73 Schultens, Albert 8, 171–82, 189–90, 193, 196,
Rijcke, Theodorus 213 200, 204, 206, 208, 211–12, 297
Rocher, Rosane 220 Schultens, Hendrik Albert 173
Rollin, Charles 211–12 Schultens, Jan Jacob 173
Roman Empire 15 Schwartz, Christian Gottlieb 125, 132–3,
patois of 212 204–6, 211
Romance languages 26, 36, 38, 68, 113, 177, 205, science communication 301–2
225, 234, 269 Scots 111
as dialects 234, 237, 239 Scythian hypothesis (Scythian as
Gallo-, see Gallo-Romance protolanguage) 162–3, 165, 219
Ibero-, see Ibero-Romance Sechehaye, Albert 245
Italo-, see Italo-Romance Séguy, Jean 276
Rhaeto- 139 semantics 5, 187, 197, 241, 301
Romansh 234 of words (lexico-) 163, 174, 178–9, 188, 190,
Romanian 139 195–6, 215, 253
Sardinian 139 as dialect-level variation 253
Romanticism 286 Semitic languages (also termed Oriental in the
and the use of dialects 286 past) 36, 38, 49, 73, 116, 141–4, 162, 165,
Ronat, Mitsou 268 172–7, 181, 183, 185, 189, 195–6,
Ronsard, Pierre de 71 199–200, 206–7, 211, 216
roofing 64, 128, 201, 271, 279 Arabic 28, 30, 49, 79, 112, 121, 141–2, 164,
low variety 201 171–3, 175–7, 179, 181, 185
roofless dialects 271 antiquity of 198
roots (of words) 9, 111–12, 142–3, 161, 163, Cairo 261
165–6, 187, 190–1, 195–6, 199, 203 dialects 190
Rosier-Catach, Irène 42 Moroccan 261
Ruhig, Philipp 144 Aramaic (Chaldean) 28, 36–7, 111–12, 115,
141, 164, 175–6, 185, 199
sacred languages 30 Biblical 30
salience 276 Galilean 39–40, 117
Salish languages 256–7 Jerusalemite 117
Sanskrit 219–20, 222–3, 226, 246 Judean 40
Hindi 227, 261 Canaanite 164
Urdu 261 Ethiopian 112, 164, 185
Sapir, Edward 248–9, 257 Hebrew 19–20, 28–30, 34, 36–7, 39–40, 49, 53,
Saporta, Sol 264 73, 79, 95, 98, 103, 111–12, 117, 121, 125,
Sappho 15 141–3, 162, 164–5, 171–6, 179–81, 185,
Sardism (Sardismós) 23–4 198–9, 205, 239
Saumaise, Claude de 8, 103–5, 107, 115–16, 166 alphabet 198
Saussure, Ferdinand de 10, 239, 244–8, 250, 273 as a dialect 162
Scaliger, Joseph Justus 7, 118, 136–41, 144, 146, dialects of 165
152–3, 196 etymology 173
Schäfer, Gottfried Heinrich 21 lexicon 198
Schleicher, August 9–10, 225–7, 229–31, 233, as the primeval language 79, 98, 119, 138,
235, 237, 242 141–3, 162, 172–3, 181
Schmeller, Johann Andreas 221 variation in 39, 98
Schmidt, Erasmus 95, 153, 198, 206 Punic 73
Schmidt, Johannes 10, 231–4, 237–8 Samaritan 112
Schnell, Rüdiger 27 Syriac 49, 73, 79, 112, 115, 121, 141, 175–6,
Schöffer, Peter 91 179, 185, 199
Schoppe, Caspar 105, 115–16 Sèrio, Luigi 201–2
Schottel, Justus Georg 97, 105, 126, 132–3, sermo 38–9, 57, 120
150, 152 Shem 79
348 
shibboleth 39–40, 165, 238–9 and nation-building 288, 293

Sinitic languages and power 288, 290, 293
Cantonese 292 and progression 290, 293
Chinese 264, 269 regionalized 277
Dongshan 261 and social class 149, 285–6
Fuzhou 261 literary 248–9
Mandarin (Standard) 292 provincial 248
politically defined 269 standard 122, 127, 214
Shanghainese 292 sub- 248
Skinner, Quentin 3 standardization 5–6, 25, 43, 50–1, 61, 69, 84–6,
slang 102, 254 88, 91–2, 106, 115, 117, 128–9, 202,
Slavic languages 37, 75, 190, 254, 272 267, 296
Bosnian 261 acceptance 50, 128, 157
Bulgarian 254 codification 50, 85, 128–9, 133, 157, 267, 271
Croatian 261 competition among languages 88
Czech (Bohemian) 37 elaboration 50–1, 85, 128–9, 133, 271, 278, 282
as one language 75, 102 and language planning 273
Polish 37 as language regulation 127
Russian 37, 227 norm 6, 50–1, 68, 125, 133, 289
Rusyn (Ruthenian) 291 processes of 130, 149, 157
as a Ukrainian dialect 291 selection 50–1, 85, 115, 117, 128, 130, 134,
Slavonic 37 247, 296
Sorbian 287 competition among dialects 61, 85
social media 293 Stanihurst, Richard 73
sociology 259, 271–4, 282, 294 Stapleton, Thomas 126
of science 109 statistics 214, 256, 260
sound changes 234 stereotyping 85
sound laws 223–4 Stevin, Simon 100
without exception (ausnahmslose Stiernhielm, Georg 7–8, 109, 112, 116–18, 144,
Lautgesetze) 228, 235, 241 151, 159–67, 176, 182, 195–6, 208, 297
speach 130 Stobaeus, Johannes 54
speech organs 285 Strabo 17, 20–1
Speroni, Sperone 50 Strodtmann, Johann Christoph 210
spiritus 161 style 5, 76, 78, 174, 197, 202, 253–4
Spitzer, Leo 235 vs. dialect 205
Sprachbund (union des langues) 253 subdialect 114, 116, 226, 236, 241
Sprache 189, 216, 225–6, 235 subdialect(us) 116–17, 153, 252, 275
Sprachlandschaft 249 substantial variation 5–7, 9–10, 59, 76, 80, 104,
Sprachmischung 234 109–24, 148, 161, 163–6, 174–6, 195, 198,
Sprachphilosophie 188, 192 203, 227, 253, 265, 271
Sprog 221 Suda 19
Sprogart 221 Suetonius 15–16, 64
Stammbaumtheorie 10, 231–3, 236 surface structure 259, 263–5
Stammwort 187 Swadesh, Morris 10, 257, 259–60, 271, 281–3
standard syllable 161, 163
colloquial 248 derivation 190
lack of 33 syntax 27, 77, 142–3, 161, 174, 178–9, 194,
language 7, 11, 71, 132–4, 149–50, 154, 250–2, 196–7, 205, 265
265, 278–80, 286, 288 as a language/dialect criterion 202
concept of 22, 127–9
and dialect 11, 132–3, 214, 247, 279, 294 Tacitus 24–5, 75
ideologies 11, 288–90, 294 Tavoni, Mirko 2
imposition of 288 terminology 238, 278–9, 282, 289, 293, 295, 300
and mutual intelligibility 292 success of 299
 349
theology 28, 214 book-length treatments of 211

Thomas, Jacques 98 in parallel with Greek dialects 88–9,
Thryllitsch, Georg Friedrich 206 155, 206
Thucydides 16 vernacular Greek (early modern) 118, 122
Tiberius, Roman emperor 15–16 dialects of 118
timelessness of language structure 253 ecclesiastical dialect 123
tópos 19, 96, 101 variation in 152
Tory, Geoffroy 88–90 vernacular languages 5, 25–6, 31, 33, 43, 50–1,
Tourtoulon, Charles de 241 61, 66, 68–71, 75, 84–6, 88, 91–2, 96,
transformation (generative concept) 263–4 106, 117, 127, 129, 139, 149, 157, 165,
translation theory 36, 42 183, 196, 250
travel 51 emancipation from Latin 85
diary 277 and nations 106
Treuge, Michael Gottlieb 206 vices of dialects 130
trilingual colleges 49 vir trilinguis 30, 49
Collège Royal (Paris) 49 Vives, Juan Luis 66, 69, 91, 128
Collegium Trilingue (Leuven) 49 Voegelin, Carl F. 10, 256–7, 262
Trissino, Gian Giorgio 99 vowel 190, 201
trópos 16 contraction 197
Trudgill, Peter 267–8 Vuidius, Robertus 194
Tryphon 21 vulgare illustre 26
túpos 19
Twitter 293 Wachter, Johann Georg 9, 200–1, 203
typophilia 241–2 Walper, Otto 125–6, 194
typophobia 241–2 WALS Online 278
Walton, Brian 122
Ulfilas 162 wanted notices 285
universal constraints 265 Warner, Levinus 171–3
ranking of 265 Weidmann, Sandra 53–4
universal language 113 Weinreich, Max 1, 249, 302
universities, medieval 28–9 Weinreich witticism 1, 158, 249–50, 268–70,
unrelated languages 187–90 282, 288, 292, 302
Uralic languages 144 attributed to Noam Chomsky 270
Urceo, Antonio, see Codro misuses of 268, 270
usage 100 Weinreich, Uriel 10, 247–51, 254, 256, 263–6,
bon usage 127 283, 298
Wenker, Georg 228, 286, 298
Valckenaer, Lodewijk Caspar 180 Whitney, William Dwight 225–8, 230
Valente, Luisa 42 Wichmann, Søren 260–1, 282
Valla, Lorenzo 26 Wilkins, John 132
Van Hal, Toon 144 Williams, John 122, 127–8
Van Keymeulen, Jacques 274, 279 Williram 185–6
Varchi, Benedetto 68, 99, 102 Wolf, Hieronymus 89
variability Wolfram, Walt 290
of human culture 42 Wood, Robert 128, 213–14
of language 17, 215–16 Wyss, Caspar 132, 208–9
variety 250, 268, 275
flexibility of 268 Yiddish 1, 249
more neutral than dialect 250, 268 YIVO (Yiddish Scientific Institute) 1
Varro, Marcus Terentius 59
Vergara, Francisco de 97 Zapotec 50
vernacular 252 Zeno 15–16
vernacular dialects 51, 73, 88–9, 128, 148, 183, zoology 52–3
207, 210–11, 216, 228, 296 Zygomalas, Theodosius 118

Language or Dialect

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Language or Dialect

Uploaded by

Copyright:

Available Formats

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 25/8/2020, SPi

RAF VAN ROOY

Some rights reserved. No part of this publication may be reproduced, stored in

I. PREHISTORY, 500 –1500

II. THE ORIGIN OF THE CONCEPTUAL

5. Lingua and dialectus: From synonymy to contrast 62

III. CONSOLIDATION BY ELABORATION,

10. The conceptual pair and language history: Language

IV. SYSTEMATIZATION AND RATIONALIZATION,

V. FROM SILENT ADOPTION TO OUTSPOKEN

23. Language, dialect, and the general public—or how not

xvi   

18.4. Hugo Schuchardt in 1922, picture by Hermann Urtel 234

4.1. Clement of Alexandria’s Greek compared to Hervet’s and Gessner’s

evolutions of the language/dialect pair in depth. Usually, it is situated in Greek

a genre of intellectual history in which episodes of contestation over meaning

embracing the upcoming standard languages. This social evolution likewise

a Byzantine author was the main source of inspiration to describe dialect-level

Chapter 21 confronts two opposite perspectives on language and the language/

16      

2.1 The Greek dialects between anecdotes and deﬁnitions

Suetonius’s anecdote about Tiberius and Zeno is a fairly banal instance of an

.    . . . 17

Dialect is speech stamped tribally and in Greek fashion, or speech of a certain

Diálektos was conceived by Diogenes as a technical term to designate features of

18      

as ‘speech that is expressible by means of letters’, phōnē ̀ eggrámmatos (ϕωνὴ

2.2 Diálektos, a variety of interpretations

Was Diogenes of Babylon alone in trying to deﬁne diálektos? Or was he part of a

⁵ Historia animalium 535a: ‘διάλεκτος δ’ ἡ τῆς ϕωνῆς ἐστι τῇ γλώττῃ διάρθρωσις’.

. ́,     19

epsilon-omicron <εο> as epsilon-upsilon <ευ>, etc. Indeed, nowhere in extant

20      

.    21

2.3 Philology as stimulus

22      

2.4 The Latin West

.    23

¹⁰ Trinummus 609: ‘ “tam modo” inquit Praenestinus’.

24      

¹² De vita Iulii Agricolae 11.3: ‘sermo haud multum diversus’.

.    25

26      

In sum, it is difﬁcult to maintain on the basis of the available evidence what

Or in a more poetical fashion:

     29

on Latin. It therefore comes as no surprise that he wanted to contribute to the

30     

3.1 Bacon at Babel

.     ? 31

3.2 Another ﬁrst for Bacon?

In his attempts at mapping out regional linguistic diversity, Bacon introduced a

For an idiom is a speciﬁc property of a language, which one people uses

32     

.     ? 33

34     

substance.¹¹ This assertion must be viewed against the background of Aristotelian

.       35

3.3 An outlook on different languages and their dialects:

36     

Here, Bacon was explaining the idiomatic characteristics of different linguistic

.       37

It is conceivable that Bacon thought Russian, Polish, and Bohemian to be idioms

38     